Depression

THE NICE GUIDELINE ON THE TREATMENT AND MANAGEMENT OF DEPRESSION IN ADULTS UPDATED EDITION

DEPRESSION
THE TREATMENT AND MANAGEMENT OF DEPRESSION IN ADULTS (UPDATED EDITION)
National Clinical Practice Guideline 90 National Collaborating Centre for Mental Health commissioned by the National Institute for Health & Clinical Excellence

published by The British Psychological Society and The Royal College of Psychiatrists

© The British Psychological Society & The Royal College of Psychiatrists, 2010
The views presented in this book do not necessarily reflect those of the British Psychological Society, and the publishers are not responsible for any error of omission or fact. The British Psychological Society is a registered charity (no. 229642). All rights reserved. No part of this book may be reprinted or reproduced or utilised in any form or by any electronic, mechanical, or other means, now known or hereafter invented, including photocopying and recording, or in any information storage or retrieval system, without permission in writing from the publishers. Enquiries in this regard should be directed to the British Psychological Society.

British Library Cataloguing-in-Publication Data A catalogue record for this book is available from the British Library. ISBN: 978-1-904671-85-5 Printed in Great Britain by Stanley Hunt. Additional material: data CD-Rom created by Pix18 (www.pix18.co.uk)
developed by National Collaborating Centre for Mental Health The Royal College of Psychiatrists 4th Floor, Standon House 21 Mansell Street London E1 8AA www.nccmh.org.uk National Institute for Health and Clinical Excellence MidCity Place, 71 High Holborn London WCIV 6NA www.nice.org.uk The British Psychological Society St Andrews House 48 Princess Road East Leicester LE1 7DR www.bps.org.uk and The Royal College of Psychiatrists 17 Belgrave Square London SW1X 8PG www.rcpsych.ac.uk

commissioned by

published by

Contents

CONTENTS
GUIDELINE DEVELOPMENT GROUP MEMBERS ACKNOWLEDGEMENTS 1 PREFACE 1.1 National guidelines 1.2 The national depression guideline 2 DEPRESSION 2.1 The disorder 2.2 Aetiology 2.3 Economic costs of depression 2.4 Treatment and management in the National Health Service 3 METHODS USED TO DEVELOP THIS GUIDELINE 3.1 Overview 3.2 The scope 3.3 The Guideline Development Group 3.4 Clinical questions 3.5 Systematic clinical literature review 3.6 Health economics methods 3.7 Methods for reviewing experience of care 3.8 Stakeholder contributions 3.9 Validation of the guideline 4 EXPERIENCE OF CARE 4.1 Introduction 4.2 Personal accounts – people with depression 4.3 Personal accounts – carers 4.4 Qualitative analysis 4.5 Review of the qualitative literature 4.6 From evidence to recommendations 4.7 Recommendations 5 CASE IDENTIFICATION AND SERVICE DELIVERY 5.1 Introduction 5.2 The identification of depression in primary care and community settings 5.3 Service delivery systems in the treatment and management of depression 5.4 Stepped care 5.5 Collaborative care 7 10 11 11 14 17 17 25 26 28 34 34 34 35 37 38 47 49 51 51 52 52 52 68 71 83 86 94 97 97 98 121 124 129 3

Contents 5.6 5.7 5.8 5.9 5.10 5.11 Medication management Crisis resolution and home treatment teams Acute day hospital care Non-acute day hospital care Non-statutory support Research recommendation 141 146 149 151 153 155 157 157 157 158 164 168 168 170 170 181 190 212 213 215 215 238 242 246 249 261 267 272 274 291 296 300 304 304 306 308 309

6 INTRODUCTION TO PSYCHOLOGICAL AND PSYCHOSOCIAL INTERVENTIONS 6.1 Introduction 6.2 Recommending psychological and psychosocial treatments 6.3 How do psychological and psychosocial interventions become evidence based? 6.4 Contextual factors that impact on clinical practice 6.5 Databases searched and inclusion/exclusion criteria 6.6 Studies considered 7 LOW-INTENSITY PSYCHOSOCIAL INTERVENTIONS 7.1 Computerised cognitive behavioural therapy 7.2 Guided self-help 7.3 Physical activity programmes 7.4 From evidence to recommendations – low-intensity psychosocial interventions 7.5 Recommendations 8 HIGH-INTENSITY PSYCHOLOGICAL INTERVENTIONS 8.1 Cognitive behavioural therapies 8.2 Behavioural activation 8.3 Problem solving 8.4 Couples therapy 8.5 Interpersonal therapy 8.6 Counselling 8.7 Short-term psychodynamic psychotherapy 8.8 Rational emotive behavioural therapy 8.9 Economic modelling 8.10 From evidence to recommendations 8.11 Recommendations 8.12 Research recommendations 9 INTRODUCTION TO PHARMACOLOGICAL AND PHYSICAL INTERVENTIONS 9.1 Introduction 9.2 Dose and duration of antidepressant treatment: evidence from clinical practice 9.3 Limitations of the literature: problems with randomised controlled trials in pharmacology 9.4 Studies considered for review – additional inclusion criteria 4

Contents 9.5 9.6 9.7 9.8 9.9 9.10 Issues and topics covered by this review Placebo-controlled randomised controlled trials of antidepressants Selective serotonin reuptake inhibitors versus placebo Tricyclic antidepressants versus placebo From evidence to recommendations Recommendation 311 313 315 319 326 327 328 328 329 330 336 341 354 354 360 387 391 398 399 411 412 413 418 418 418 424 425 427 430 451 453 457 462

10 PHARMACOLOGICAL INTERVENTIONS 10.1 Introduction 10.2 Use of individual drugs in the treatment of depression 10.3 Tricyclic antidepressants 10.4 Selective serotonin reuptake inhibitors 10.5 Escitalopram 10.6 The thread study 10.7 Monoamine oxidase inhibitors 10.8 Third-generation antidepressants 10.9 St John’s wort 10.10 Health economics evidence 10.11 Network meta-analysis of newer antidepressants 10.12 Economic model for the cost-effectiveness of pharmacological interventions for people with depression 10.13 From evidence to recommendations 10.14 Clinical practice recommendations 10.15 When to change antidepressant treatment when symptoms of depression are not improving 11 FACTORS INFLUENCING CHOICE OF ANTIDEPRESSANTS 11.1 Introduction 11.2 The pharmacological management of depression in older adults 11.3 The effect of sex on antidepressant choice 11.4 The pharmacological management of depression with psychotic symptoms 11.5 The pharmacological management of atypical depression 11.6 The physical and pharmacological management of depression with a seasonal pattern 11.7 Dosage issues for tricyclic antidepressants 11.8 Antidepressant discontinuation symptoms 11.9 The cardiotoxicity of antidepressants 11.10 Depression, antidepressants and suicide 12 THE PHARMACOLOGICAL AND PHYSICAL MANAGEMENT OF DEPRESSION THAT HAS NOT ADEQUATELY RESPONDED TO TREATMENT, AND RELAPSE PREVENTION 12.1 Introduction 12.2 Approach to the reviews 12.3 Pharmacological ‘next-step’ treatment for depression that has not adequately responded to treatment

466 466 467 469 5

Contents 12.4 Electroconvulsive therapy 12.5 Other non-pharmacological physical treatments 12.6 The pharmacological management of relapse prevention 13 THE MANAGEMENT OF SUBTHRESHOLD DEPRESSIVE SYMPTOMS 13.1 Introduction 13.2 Pharmacological interventions for subthreshold depressive symptoms and persistent subthreshold depressive symptoms (dysthymia) 13.3 Psychological and other strategies for the treatment of persistent subthreshold depressive symptoms (dysthymia) 13.4 From evidence to recommendations 13.5 Recommendations 13.6 Research recommendation 14 SUMMARY OF RECOMMENDATIONS 14.1 Care of all people with depression 14.2 Stepped care 14.3 Step 1: recognition, assessment and initial management 14.4 Step 2: recognised depression – persistent subthreshold depressive symptoms or mild to moderate depression 14.5 Step 3: persistent subthreshold depressive symptoms or mild to moderate depression with inadequate response to initial interventions, and moderate and severe depression 14.6 Treatment choice based on depression subtypes and personal characteristics 14.7 Enhanced care for depression 14.8 Sequencing treatments after initial inadequate response 14.9 Continuation and relapse prevention 14.10 Step 4: complex and severe depression 14.11 Research recommendations 15 APPENDICES 16 REFERENCES 17 ABBREVIATIONS 508 528 530 536 536

537 555 563 564 564 565 565 567 568 569

571 576 576 577 579 582 585 591 647 696

6

Guideline Development Group members

GUIDELINE DEVELOPMENT GROUP MEMBERS
Professor Ian Anderson (Chair, Guideline Development Group) Professor of Psychiatry, University of Manchester Professor Stephen Pilling Director, National Collaborating Centre for Mental Health Director, Centre for Outcomes Research and Effectiveness, University College London Ms Alison Barnes Service User Member Ms Linda Bayliss Research Assistant (May 2008 to August 2008), National Collaborating Centre for Mental Health Ms Victoria Bird Research Assistant, National Collaborating Centre for Mental Health Ms Rachel Burbeck Lead Systematic Reviewer, National Collaborating Centre for Mental Health Dr Carolyn Chew-Graham General Practitioner and Senior Lecturer in Primary Care, University of Manchester Mr Jeremy Clarke Psychological Therapist, Lambeth Primary Care Trust Mr Matthew Dyer Health Economist, National Collaborating Centre for Mental Health Ms Esther Flanagan Project Manager (2009), National Collaborating Centre for Mental Health Ms Catherine Harris Carer member and Local Councillor Ms Sarah Hopkins Project Manager (until 2008), National Collaborating Centre for Mental Health

7

Guideline Development Group members Dr Mark Kenwright Consultant Cognitive Behavioural Psychotherapist, Ealing Cognitive Behavioural Therapy Service Professor Willem Kuyken Professor of Clinical Psychology and Co-Director, Mood Disorders Centre, School of Psychology, University of Exeter Ms Angela Lewis Research Assistant, National Collaborating Centre for Mental Health Professor Glyn Lewis Professor of Psychiatric Epidemiology, University of Bristol Mr Ryan Li Project Manager (2008), National Collaborating Centre for Mental Health Mr Brendan Masterson Clinical Nurse Leader, Affective Disorders Unit, Bethlem Royal Hospital Dr Nick Meader Systematic Reviewer, National Collaborating Centre for Mental Health Mr Alan Meudell Service User Member, Healthy Minds at Work Dr Alex Mitchell Consultant Psychiatrist and Honorary Lecturer in Liaison Psychiatry, University of Leicester Dr Richard Moore Clinical Psychologist, Cambridge and Peterborough NHS Foundation Trust Dr Suffiya Omarjee Health Economist, National Collaborating Centre for Mental Health Ms Carol Paton Chief Pharmacist, Oxleas NHS Foundation Trust Dr Alejandra Perez Systematic Reviewer, National Collaborating Centre for Mental Health Ms Peny Retsa Health Economist (until 2008), National Collaborating Centre for Mental Health

8

Guideline Development Group members Ms Maria Rizzo Research Assistant, National Collaborating Centre for Mental Health Ms Jennie Robertson Research Assistant (from September 2008), National Collaborating Centre for Mental Health Mr Rob Saunders Research Assistant (2008), National Collaborating Centre for Mental Health Ms Christine Sealey Centre Manager, National Collaborating Centre for Mental Health Ms Beth Shackleton Project Manager (until 2008), National Collaborating Centre for Mental Health Dr Thomas Shackleton General Practitioner, Suffolk Ms Sarah Stockton Senior Information Scientist, National Collaborating Centre for Mental Health Dr Clare Taylor Editor, National Collaborating Centre for Mental Health Ms Jane Wood Nurse, Strategic Development Manager, Mental Health, Leeds Primary Care Trust

9

Acknowledgements

ACKNOWLEDGEMENTS
Editorial assistance Ms Nuala Ernest Ms Marie Halton

10

Preface

1

PREFACE

This guideline was first published in December 2004 (NICE, 2004a; NCCMH, 2004) (referred to as the ‘previous guideline’). The present guideline (referred to as the ‘update’) updates many areas of the previous guideline. There are also new chapters on the experience of depression for people with depression and their carers (Chapter 4), and on the treatment and management of subthreshold depressive symptoms (including dysthymia symptoms) (Chapter 13), which were not part of the scope of the previous guideline. Recommendations categorised as ‘good practice points’ in the previous guideline were reviewed for their current relevance (including issues around consent and advance directives). Further details of what has been updated and what is left unchanged can be found at the beginning of each evidence chapter. The scope for the update also included updating two National Institute for Health and Clinical Excellence (NICE) technology appraisals (TAs) on the use of electroconvulsive therapy (ECT) (TA59) and on computerised cognitive behaviour therapy (TA51) (NICE, 2003, 2002)1. See Appendix 1 for more details on the scope of this update. Sections of the guideline where the evidence has not been updated are marked by asterisks (**_**). The previous guideline and this update have been developed to advise on the treatment and management of depression. The guideline recommendations in the update have been developed by a multidisciplinary team of healthcare professionals, people with depression, a carer and guideline methodologists after careful consideration of the best available evidence. It is intended that the guideline will be useful to clinicians and service commissioners in providing and planning high-quality care for people with depression while also emphasising the importance of the experience of care for them and their carers. Although the evidence base is rapidly expanding there are a number of major gaps, and further revisions of this guideline will incorporate new scientific evidence as it develops. The guideline makes a number of research recommendations specifically to address gaps in the evidence base. In the meantime, it is hoped that the guideline will assist clinicians, people with depression and their carers by identifying the merits of particular treatment approaches where the evidence from research and clinical experience exists.

1.1 1.1.1

NATIONAL GUIDELINES What are clinical practice guidelines?

Clinical practice guidelines are ‘systematically developed statements that assist clinicians and patients in making decisions about appropriate treatment for specific condifrom TA59 and TA97 were incorporated into the previous depression guideline according to NICE protocol.
1Recommendations

11

Preface tions’ (Mann, 1996). They are derived from the best available research evidence, using predetermined and systematic methods to identify and evaluate the evidence relating to the specific condition in question. Where evidence is lacking, the guidelines incorporate statements and recommendations based upon the consensus statements developed by the Guideline Development Group (GDG). Clinical guidelines are intended to improve the process and outcomes of healthcare in a number of different ways. They can: ● provide up-to-date evidence-based recommendations for the management of conditions and disorders by healthcare professionals ● be used as the basis to set standards to assess the practice of healthcare professionals ● form the basis for education and training of healthcare professionals ● assist people with depression and their carers in making informed decisions about their treatment and care ● improve communication between healthcare professionals, people with depression and their carers ● help identify priority areas for further research.

1.1.2

Uses and limitations of clinical guidelines

Guidelines are not a substitute for professional knowledge and clinical judgement. They can be limited in their usefulness and applicability by a number of different factors: the availability of high-quality research evidence, the quality of the methodology used in the development of the guideline, the generalisability of research findings and the uniqueness of individuals with depression. Although the quality of research in this field is variable, the methodology used here reflects current international understanding on the appropriate practice for guideline development (AGREE: Appraisal of Guidelines for Research and Evaluation Instrument; www.agreetrust.org; AGREE Collaboration [2003]), ensuring the collection and selection of the best research evidence available and the systematic generation of treatment recommendations applicable to the majority of people with depression. However, there will always be some people and situations for which clinical guideline recommendations are not readily applicable. This guideline does not, therefore, override the individual responsibility of healthcare professionals to make appropriate decisions in the circumstances of the individual, in consultation with the person with depression or their carer. In addition to the clinical evidence, cost-effectiveness information, where available, is taken into account in the generation of statements and recommendations in clinical guidelines. While national guidelines are concerned with clinical and cost effectiveness, issues of affordability and implementation costs are to be determined by the National Health Service (NHS). In using guidelines, it is important to remember that the absence of empirical evidence for the effectiveness of a particular intervention is not the same as evidence for ineffectiveness. In addition, of particular relevance in mental health, evidencebased treatments are often delivered within the context of an overall treatment 12

Preface programme including a range of activities, the purpose of which may be to help engage the person and to provide an appropriate context for the delivery of specific interventions. It is important to maintain and enhance the service context in which these interventions are delivered; otherwise the specific benefits of effective interventions will be lost. Indeed, the importance of organising care in order to support and encourage a good therapeutic relationship is at times as important as the specific treatments offered.

1.1.3

Why develop national guidelines?

NICE was established as a Special Health Authority for England and Wales in 1999, with a remit to provide a single source of authoritative and reliable guidance for patients, professionals and the public. NICE guidance aims to improve standards of care, diminish unacceptable variations in the provision and quality of care across the NHS and ensure that the health service is patient centred. All guidance is developed in a transparent and collaborative manner using the best available evidence and involving all relevant stakeholders. NICE generates guidance in a number of different ways, three of which are relevant here. First, national guidance is produced by the Technology Appraisal Committee to give robust advice about a particular treatment, intervention, procedure or other health technology. Second, NICE commissions public health intervention guidance focused on types of activity (interventions) that help to reduce people’s risk of developing a disease or condition or help to promote or maintain a healthy lifestyle. Third, NICE commissions the production of national clinical practice guidelines focused upon the overall treatment and management of a specific condition. To enable this latter development, NICE originally established seven National Collaborating Centres in conjunction with a range of professional organisations involved in healthcare.

1.1.4

The National Collaborating Centre for Mental Health

This guideline has been commissioned by NICE and developed within the National Collaborating Centre for Mental Health (NCCMH). The NCCMH is a collaboration of the professional organisations involved in the field of mental health, national patient and carer organisations, and a number of academic institutions and NICE. The NCCMH is funded by NICE and is led by a partnership between the Royal College of Psychiatrists and the British Psychological Society’s Centre for Outcomes Research and Effectiveness.

1.1.5

From national guidelines to local implementation

Once a national guideline has been published and disseminated, local healthcare groups will be expected to produce a plan and identify resources for implementation, 13

Preface along with appropriate timetables. Subsequently, a multidisciplinary group involving commissioners of healthcare, primary care, specialist mental health professionals, and people with depression and their carers should undertake the translation of the implementation plan locally, taking into account both the recommendations set out in this guideline and the priorities set in the National Service Framework for Mental Health (Department of Health, 1999) and related documentation. The nature and pace of the local plan will reflect local healthcare needs and the nature of existing services; full implementation may take considerable time, especially where substantial training needs are identified.

1.1.6

Auditing the implementation of guidelines

This guideline identifies key areas of clinical practice and service delivery for local and national audit. Although the generation of audit standards is an important and necessary step in the implementation of this guidance, a more broadly based implementation strategy will be developed. Nevertheless, it should be noted that the Healthcare Commission will monitor the extent to which Primary Care Trusts, trusts responsible for mental health and social care and Health Authorities have implemented these guidelines.

1.2 1.2.1

THE NATIONAL DEPRESSION GUIDELINE Who has developed this guideline?

The GDG was convened by the NCCMH and supported by funding from NICE. The GDG included two people with depression and a carer, and professionals from psychiatry, clinical psychology, general practice, nursing and psychiatric pharmacy. Staff from the NCCMH provided leadership and support throughout the process of guideline development, undertaking systematic searches, information retrieval, appraisal and systematic review of the evidence. Members of the GDG received training in the process of guideline development from NCCMH staff, and the people with depression and the carer received training and support from the NICE Patient and Public Involvement Programme. The NICE Guidelines Technical Adviser provided advice and assistance regarding aspects of the guideline development process. All GDG members made formal declarations of interest at the outset, which were updated at every GDG meeting. The GDG met a total of 14 times throughout the process of guideline development. It met as a whole, but key topics were led by a national expert in the relevant topic. The GDG was supported by the NCCMH technical team, with additional expert advice from special advisers where needed. The group oversaw the production and synthesis of research evidence before presentation. All statements and recommendations in this guideline have been generated and agreed by the whole GDG. 14

Preface 1.2.2 For whom is this guideline intended?

This guideline is relevant for adults with depression as the primary diagnosis and covers the care provided by primary, community, secondary, tertiary and other healthcare professionals who have direct contact with, and make decisions concerning the care of, adults with depression. The guideline will also be relevant to the work, but will not cover the practice, of those in: ● occupational health services ● social services ● forensic services ● the independent sector. The experience of depression can affect the whole family and often the community. The guideline recognises the role of both in the treatment and support of people with depression.

1.2.3

Specific aims of this guideline

The guideline makes recommendations for the treatment and management of depression. It aims to: ● improve access and engagement with treatment and services for people with depression ● evaluate the role of specific psychological and psychosocial interventions in the treatment of depression ● evaluate the role of specific pharmacological interventions in the treatment of depression ● evaluate the role of specific service-level interventions for people with depression ● integrate the above to provide best-practice advice on the care of people with depression and their family and carers ● promote the implementation of best clinical practice through the development of recommendations tailored to the requirements of the NHS in England and Wales.

1.2.4

The structure of this guideline

The guideline is divided into chapters, each covering a set of related topics. The first three chapters provide an introduction to guidelines, the topic of depression and the methods used to update this guideline. Chapters 5 to 13 provide the evidence that underpins the recommendations about the treatment and management of depression, with Chapter 4 providing personal accounts from people with depression and carers that offer an insight into their experience of depression. Each evidence chapter begins with a general introduction to the topic that sets the recommendations in context. Depending on the nature of the evidence, narrative reviews or meta-analyses were conducted, and the structure of the chapters varies 15

Preface accordingly. Where appropriate, details about current practice, the evidence base and any research limitations are provided. Where meta-analyses were conducted, information is given about the review protocol and studies included in the review. Clinical evidence summaries are used to summarise the data presented. Health economic evidence is then presented (where appropriate), followed by a section (from evidence to recommendations) that draws together the clinical and health economic evidence and provides a rationale for the recommendations. On the CD-ROM, further details are provided about included/excluded studies, the evidence, and the previous guideline methodology (see Table 1 for details).

Table 1: Appendices on CD-ROM Evidence tables for economic studies Clinical evidence profiles Clinical study characteristics tables References to studies from the previous guideline Clinical evidence forest plots Case identification included and excluded studies Previous guideline methodology Appendix 15 Appendix 16 Appendix 17 Appendix 18 Appendix 19 Appendix 20 Appendix 21

16

Depression

2

DEPRESSION

This guideline is concerned with the treatment and management of adults with a primary diagnosis of depression in primary and secondary care. The terminology and diagnostic criteria used for this heterogeneous group of related disorders have changed over the years, and the previous guideline related only to those identified by The ICD–10 Classification of Mental and Behavioural Disorders (ICD–10) (WHO, 1992) as having a depressive episode (F32 in the ICD–10), recurrent depressive episode (F33) or mixed anxiety and depressive disorder (F41.2). In this guideline update the scope was widened to cover the substantial proportion of people who present with less severe forms of depression. Therefore, this updated guideline covers ‘subthreshold depressive symptoms’, which fall below the criteria for major depression (and which do not have a coding in ICD–10), and subthreshold depressive symptoms persisting for at least 2 years (dysthymia; F34.1). It should, however, be noted that much of the research forming the evidence base from which this guideline is drawn has used a different classificatory system – the Diagnostic and Statistical Manual of Mental Disorders of the American Psychiatric Association, currently in its fourth edition (DSM–IV-TR) (APA, 2000c). The two classificatory systems, while similar, are not identical especially with regard to definitions of severity. After considerable discussion the GDG took the decision to base the guidelines on the DSM–IV-TR (see Section 2.1.5). This covers major depressive disorder single episode (296.2) and recurrent (296.3) together with dysthymic disorder (300.4), and contains research criteria for minor depressive disorder (APA, 2000c). The effect of this change in practice is discussed in Section 2.1.5 (see also Appendix 11). The guideline does not address the management of depression in children and adolescents, depression in bipolar disorder, depression occurring in both antenatal and postnatal periods, or depression associated with chronic physical health problems, all of which are covered by separate guidelines (NICE, 2005, 2006c, 2007e, 2009c). The guideline update does cover psychotic symptoms occurring within the context of an episode of depression (depression with psychotic symptoms), but not depression occurring in a primary psychotic illness, such as schizophrenia or dementia.

2.1 2.1.1

THE DISORDER Symptoms, presentation and pattern of illness

Depression refers to a wide range of mental health problems characterised by the absence of a positive affect (a loss of interest and enjoyment in ordinary things and experiences), low mood and a range of associated emotional, cognitive, physical and behavioural symptoms. Distinguishing the mood changes between clinically significant 17

Depression degrees of depression (for example, major depression) and those occurring ‘normally’ remains problematic and it is best to consider the symptoms of depression as occurring on a continuum of severity (Lewinsohn et al., 2000). The identification of major depression is based not only on its severity but also on persistence, the presence of other symptoms, and the degree of functional and social impairment. However, there appears to be no hard-and-fast ‘cut-off’ between ‘clinically significant’ and ‘normal’ degrees of depression; the greater the severity of depression, the greater the morbidity and adverse consequences (Lewinsohn et al., 2000; Kessing, 2007). When taken together with other aspects that need to be considered, such as duration, stage of illness and treatment history, there are considerable problems when attempting to classify depression into categories (see Section 2.1.5). Commonly, mood and affect in a major depressive illness are unreactive to circumstance, remaining low throughout the course of each day, although for some people mood varies diurnally, with gradual improvement throughout the day only to return to a low mood on waking. For others, a person’s mood may be reactive to positive experiences and events, although these elevations in mood are not sustained, with depressive feelings re-emerging, often quickly (Andrews & Jenkins, 1999). Behavioural and physical symptoms typically include tearfulness, irritability, social withdrawal, an exacerbation of pre-existing pains, pains secondary to increased muscle tension (Gerber et al., 1992), a lack of libido, fatigue and diminished activity, although agitation is common and marked anxiety frequent. Typically there is reduced sleep and lowered appetite (sometimes leading to significant weight loss), but for some people it is recognised that sleep and appetite are increased. A loss of interest and enjoyment in everyday life, and feelings of guilt, worthlessness and that one deserves punishment, are common, as are lowered self-esteem, loss of confidence, feelings of helplessness, suicidal ideation and attempts at self-harm or suicide. Cognitive changes include poor concentration and reduced attention, pessimistic and recurrently negative thoughts about oneself, one’s past and the future, mental slowing and rumination (Cassano & Fava, 2002). Depression is often accompanied by anxiety, and in these circumstances one of three diagnoses can be made: (1) depression; (2) anxiety; or (3) mixed depression and anxiety when both are below the threshold for either disorder, dependent upon which constellation of symptoms dominates the clinical picture. In addition, the presentation of depression can vary with age with the young showing more behavioural symptoms and older adults more somatic symptoms and fewer complaints of low mood (Serby & Yu, 2003). Major depression is generally diagnosed when a persistent low mood and an absence of positive affect are accompanied by a range of symptoms, the number and combination needed to make a diagnosis being operationally defined (ICD–10, WHO, 1992; DSM–IV, APA, 1994). Some people are recognised as showing an atypical presentation with reactive mood, increased appetite, weight gain and excessive sleepiness together with the personality feature of sensitivity to rejection (Quitkin et al., 1991) and this is classified as major depression with atypical features in DSM–IV (APA, 1994). The definition of atypical depression has changed over time and it is not specifically recognised in ICD–10. 18

Depression Some patients have a more severe and typical presentation, including marked physical slowness (or marked agitation), complete lack of reactivity of mood to positive events, and a range of somatic symptoms, including appetite and weight loss, reduced sleep with a particular pattern of waking early in the morning and being unable to get back to sleep. A pattern of the depression being substantially worse in the morning (diurnal variation) is also commonly seen. This presentation is referred to as major depression with melancholic features in DSM–IV and a depressive episode with somatic symptoms in ICD–10. People with severe depression may also develop psychotic symptoms (hallucinations and/or delusions), most commonly thematically consistent with the negative, self-blaming cognitions and low mood typically encountered in major depression, although others may develop psychotic symptoms unrelated to mood (Andrews & Jenkins, 1999). In the latter case, these mood-incongruent psychotic symptoms can be hard to distinguish from those that occur in other psychoses such as schizophrenia.

2.1.2

Course and prognosis

The average age of the first episode of major depression occurs in the mid-20s and, although the first episode may occur at any time from early childhood through to old age, a substantial proportion of people have their first depression in childhood or adolescence (Fava & Kendler, 2000). Just as the initial presentation and form of a depressive illness varies considerably, so too does the prodromal period. Some individuals experience a range of symptoms in the months prior to the full illness, including anxiety, phobias, milder depressive symptoms and panic attacks; others may develop a severe major depressive illness fairly rapidly, not uncommonly following a major stressful life event. Sometimes somatic symptoms dominate the clinical picture leading the clinician to investigate possible underlying physical illness until mood changes become more obvious. Although depression has been thought of as a time-limited disorder, lasting on average 4 to 6 months with complete recovery afterwards, it is now clear that incomplete recovery and relapse are common. The WHO study of mental disorders in 14 centres across the world found that 50% of patients still had a diagnosis of depression 1 year later (Simon et al., 2002) and at least 10% had persistent or chronic depression (Kessler et al., 2003). At least 50% of people, following their first episode of major depression, will go on to have at least one more episode (Kupfer, 1991) and, after the second and third episodes, the risk of further relapse rises to 70 and 90%, respectively (Kupfer, 1991). People with early onset depression (at or before 20 years of age) and depression occurring in old age have a significantly increased vulnerability to relapse (Giles et al., 1989; Mitchell & Subramaniam, 2005). Thus, while the outlook for a first episode is good, the outlook for recurrent episodes over the long term can be poor with many patients experiencing symptoms of depression over many years (Akiskal, 1986). Sometimes, recurrent episodes of depression will follow a seasonal pattern which has been called ‘seasonal affective disorder’ (SAD; Rosenthal et al., 1984). DSM–IV includes 19

Depression criteria for a seasonal pattern whereas only provisional criteria are given in the research version of ICD–10. Although a seasonal pattern can apply to both recurrent depression and bipolar disorder it appears most common in the former (70 to 80%, Rodin & Thompson, 1997; Westrin & Lam, 2007), with recurrent winter depression far more common than recurrent summer episodes (Rodin & Thompson, 1997; Magnusson & Partonen, 2005). Depression with a seasonal pattern refers to depression that occurs repeatedly at the same time of year (not accounted for by psychosocial stress) with remission in between and without a lifetime predominance of non-seasonal depression. Decreased activity is reported as nearly always present and atypical depressive symptoms, particularly increased sleep, weight gain and carbohydrate craving are common (Magnusson & Partonen, 2005). The onset is reported as usually in the third decade and is more common in the young (Rodin & Thompson, 1997; Magnusson & Partonen, 2005). Surveys in the UK have found a surprisingly high prevalence in general practitioner (GP) practice attendees ranging from 3.5% in Aberdeen (Eagles et al., 1999) to 5.6% in southern England (Thompson et al., 2004). However, the validity of ‘seasonal affective disorder’ has been poorly accepted in Europe and may be an extreme form of a dimensional ‘seasonality trait’ rather than a specific diagnosis (Kasper et al., 1989). Some patients with non-seasonal mood disorders also report seasonal variation (Bauer & Dunner, 1993) and this also occurs in other disorders such as anxiety and eating disorders (Bauer & Dunner, 1993; Magnusson & Partonen, 2005). After 5 to 11 years’ follow-up, approximately half of those with continuing depressive episodes no longer display a seasonal pattern (Magnusson & Partonen, 2005). Up to 10% of people with depression subsequently experience hypomanic/manic episodes (Kovacs, 1996), which emphasises the need to question patients about a history of elevated mood and to be alert to new episodes occurring. In the WHO study, episodes of depression that were either untreated by the GP or missed entirely had the same outlook as treated episodes of depression; however, they were milder at index consultation (Goldberg et al., 1998). A small longitudinal study (Kessler et al., 2002) found that the majority of undetected people either recovered or were diagnosed during the follow-up period; nevertheless, nearly 20% of the identified cases in this study remained undetected and unwell after 3 years. The term ‘treatment-resistant depression’ was used in the previous guideline to describe depression that has failed to respond to two or more antidepressants at an adequate dose for an adequate duration given sequentially. Although the term is commonly used, and it can be seen as a useful ‘short-hand’ to refer to difficulties in achieving adequate improvement with treatment, it has problems that led the GDG to a move away from its use in this guideline update. The term implies that there is a natural cut-off between people who respond to one or two antidepressants compared with those who do not, which is not supported by the evidence, and the term may be taken by both doctors and patients as a pejorative label. It is also not helpful as it does not take into account different degrees of improvement or stages of illness (whether occurring in an ongoing episode or relapse in spite of ongoing treatment). It takes no account of psychotherapeutic treatment, and non-antidepressant augmenting agents are not easily incorporated. The limited trial evidence base reflects the lack of a natural distinction and different studies incorporate different degrees of treatment failure. Finally, it fails to take 20

Depression into account whether psychosocial factors may be preventing recovery (Andrews & Jenkins, 1999). The GDG preferred to approach the problem of inadequate response by considering sequenced treatment options rather than by a category of patient.

2.1.3

Disability and mortality

Depression is the most common mental disorder in community settings and is a major cause of disability across the world. In 1990 it was the fourth most common cause of loss of disability-adjusted life years (DALYs) in the world, and it is projected to become the second most common cause by 2020 (World Bank, 1993). In 1994, it was estimated that about 1.5 million DALYs were lost each year in the West as a result of depression (Murray et al., 1994). It is even more common in the developing world (for a review, see Institute of Medicine, 2001). There is a clear dose–response relationship between illness severity and the extent of disability (Ormel & Costa e Silva, 1995) and onsets of depression are associated with onsets of disability, with an approximate doubling of both social and occupational disability (Ormel et al., 1999). Apart from the subjective experiences of people with depression, the impact on social and occupational functioning, physical health and mortality is substantial. Depressive illness causes a greater decrement in health state than the major chronic physical illnesses: angina, arthritis, asthma and diabetes (Moussavi et al., 2007). Emotional, motivational and cognitive effects substantially reduce a person’s ability to work effectively, with losses in personal and family income as well as lost contribution to society in tax revenues and employment skills. Wider social effects include: greater dependence upon welfare and benefits, with loss of self-esteem and self-confidence; social impairments, including reduced ability to communicate and sustain relationships during the illness with knock-on effects after an episode; and longer-term impairment in social functioning, especially for those who have chronic or recurrent disorders. The stigma associated with mental health problems generally (Sartorius, 2002), and the public view that others might view a person with depression as unbalanced, neurotic and irritating (Priest et al., 1996), may partly account for the reluctance of people with depression to seek help (Bridges & Goldberg, 1987). Depression can also exacerbate the pain, distress and disability associated with physical health problems as well as adversely affecting outcomes. Depression combined with chronic physical health problems incrementally worsens health compared with physical disease alone or even combinations of physical diseases (Moussavi et al., 2007). In addition, for a range of physical health problems, findings suggest an increased risk of death when comorbid depression is present (Cassano & Fava, 2002). In coronary heart disease, for example, depressive disorders are associated with an 80% increased risk, both of its development and of subsequent mortality in established disease, at least partly through common contributory factors (Nicholson et al., 2006). Another guideline on depression in adults with a chronic physical health problem accompanies this guideline update (NCCMH, 2010). Suicide accounts for nearly 1% of all deaths and nearly two-thirds of this figure occur in people with depression (Sartorius, 2001). Looked at another way, having 21

Depression depression leads to over a four-times higher risk of suicide compared with the general population, which rises to nearly 20 times in the most severely ill (Bostwick & Pankratz, 2000). Sometimes depression may also lead to acts of violence against others and may even include homicide. Marital and family relationships are frequently negatively affected, and parental depression may lead to neglect of children and significant disturbances in children (Ramachandani & Stein, 2003).

2.1.4

Incidence and prevalence

Worldwide estimates of the proportion of people who are likely to experience depression in their lifetime vary widely between studies and settings, but the best estimates lie between about 4 and 10% for major depression, and between about 2.5 and 5% for dysthymia (low grade chronic depressive symptoms) (Waraich et al., 2004) with disparities attributable to real differences between countries and the method of assessment. The estimated point prevalence for a depressive episode (F32/33, ICD–10; WHO, 1992) among 16- to 74-year-olds in the UK in 2000 was 2.6% (males 2.3%, females 2.8%), but, if the broader and less specific category of ‘mixed depression and anxiety’ (F41.2, ICD–10, WHO, 1992) was included, these figures rose dramatically to 11.4% (males 9.1%, females 13.6%) (Singleton et al., 2001). Prevalence rates have consistently been found to be between 1.5 and 2.5 times higher in women than men and have also been fairly stable in the age range of 18 to 64 years (Waraich et al., 2004), although in the most recent UK survey cited above female preponderance was only marked for a depressive episode in those under 35 years whereas for mixed anxiety and depression it was across the age range. Compared with adults without a neurotic disorder, those with a depressive episode or mixed anxiety and depression were more likely to be aged between 35 and 54 years, separated or divorced and living alone or as a lone parent. This pattern was broadly similar between men and women (Singleton et al., 2001). A number of socioeconomic factors significantly affected prevalence rates in the UK survey: those with a depressive episode were more likely than those without ‘neurotic disorders’ (depressive or anxiety disorders) to be unemployed, to belong to social classes 4 and below, to have lower predicted intellectual function, to have no formal educational qualifications and to live in local authority or Housing Association accommodation, to have moved three or more times in the last 2 years and to live in an urban environment (Singleton et al., 2001). No significant effect of ethnic status on prevalence rates of a depressive episode or mixed anxiety and depression were found, although numerically there was a higher proportion of South Asians in those with depressive or anxiety disorders than in those without (Singleton et al., 2001). Migration has been high in Europe in the last 2 decades, but data on mental health is scarce and results vary between migrant groups (Lindert et al., 2008). An illustration of the social origins of depression can be found in a general practice survey in which 7.2% (range 2.4 to 13.7%, depending upon the practice) of consecutive attendees had a depressive disorder. Neighbourhood social deprivation 22

there is no natural discontinuity between subthreshold depressive symptoms and ‘mild major’ depression in routine clinical practice. This may mean that more people may be identified as depressed using ICD–10 criteria compared with DSM–IV (Wittchen et al. Although the advent of operational diagnostic criteria has improved the reliability of diagnosis. varies according to gender. 2. 2001a). and that 23 . Diagnostic criteria and methods of classification of depressive disorders have changed substantially over the years. have virtually the same diagnostic features for a ‘clinically important’ severity of depression (termed a major depressive episode in DSM–IV or a depressive episode in ICD–10). At least two but less than five symptoms are required and it overlaps with ICD–10 mild depressive episode with four symptoms. These studies emphasise that. related to the need for only one of two key symptoms for DSM–IV but two out of three for ICD–10.Depression accounted for 48.5 Diagnosis In recent years there has been a greater recognition of the need to consider depression that is ‘subthreshold’. The GDG has widened the range of depressive disorders to be considered in this guideline update and emphasises that the diagnostic ‘groupings’ it uses should be viewed as pragmatic subdivisions of dimensions in the form of vignettes or exemplars rather than firm categories. The evidence therefore overwhelmingly supports the view that the prevalence of depression. anhedonia and loss of energy). The GDG considered it important to acknowledge the uncertainty inherent in our current understanding of depression and its classification. and are more common in those with a history of major depression as well as being a risk factor for future major depression (Rowe & Rapaport. the two systems are not identical and that this is particularly apparent at the threshold taken to indicate clinical importance. where the depression does not meet the full criteria for a depressive/major depressive episode. that is.1.. see Appendix 11). or at least that somewhat different populations are identified (Andrews et al. 2001). and neighbourhood unemployment (Ostler et al. Nevertheless their thresholds differ. this does not circumvent the fundamental problem of attempting to classify a disorder that is heterogeneous and best considered in a number of dimensions (for a fuller discussion. There is no accepted classification for subthreshold depression in the current diagnostic systems.. DSM–IV and ICD–10. however it is defined. 2006). with DSM–IV requiring a minimum of five out of nine symptoms (which must include depressed mood and/or anhedonia) and ICD–10 requiring four out of ten symptoms (including at least two of depressed mood. Subthreshold depressive symptoms cause considerable morbidity and human and economic costs. and social and economic factors. with the closest being minor depression (a research diagnosis in DSM–IV). although similar..3% of the variance among practices and the variables that accounted for most of that variance were: the proportion of the population having no or only one car. Given the practical difficulty and inherent uncertainty in deciding thresholds for significant symptom severity and disability. 2008).

if any. adapted from DSM–IV. Gender and socioeconomic factors account for large variations in the population rates of depression and few studies of pharmacological. Diagnosis requires a minimum of 2 weeks’ duration of symptoms that includes at least one key symptom. damaging. to emphasise that symptom severity rating scales should not be used by themselves to make the diagnosis. and the symptoms markedly interfere with functioning. It is also common for depressed people to have a comorbid psychiatric diagnosis. diagnosis using the three factors listed above (severity. although they can be an aid in assessing severity and response to treatment. duration and course) only provides a partial description of the individual experience of depression. the GDG for the update used DSM–IV rather than ICD–10 to define the diagnosis of depression because the evidence base for treatments nearly always uses DSM–IV. their family history. reinforcing the need to be circumspect about an over-rigid extrapolation from RCTs to clinical practice. Patients seen clinically are rarely assessed using standardised criteria. psychological or indeed other treatments for depression either 24 . The implication of the change in diagnostic system used in the guideline update. sexual abuse). To make a diagnosis of a depression requires assessment of three linked but separate factors: (a) severity. (b) duration and (c) course. such as anxiety. A diagnosis is a starting point in considering the most appropriate way of helping that individual in their particular circumstances. Can occur with or without psychotic symptoms. is that it is likely to raise the thresholds for some specific treatments such as antidepressants. psychological mindedness and current relational and social problems – all of which may significantly affect outcomes. combined with redefining the severity ranges. and be present for most of every day. the GDG attempted to move away from focusing on one aspect such as severity. An important motivation has been to provide a strong steer away from only using symptom counting to make the diagnosis of depression and. are used in the guideline update: ● subthreshold depressive symptoms: fewer than five symptoms of depression ● mild depression: few. which can have the unwanted effect of leading to the categorisation of depression and influencing treatment choice based on a single factor such as a symptom count. social phobia. personalities. by extension. The following definitions of depression. even worse. and the symptoms result in only minor functional impairment ● moderate depression: symptoms or functional impairment are between ‘mild’ and ‘severe’ ● severe depression: most symptoms.. In addition. premorbid difficulties (for example. However. Individual symptoms should be assessed for severity and impact on function.Depression assuming a false categorical certainty is likely to be unhelpful and. People with depression vary in the pattern of symptoms they experience. symptoms in excess of the five required to make the diagnosis. It is important to emphasise that making a diagnosis of depression does not automatically imply a specific treatment. and physical comorbidity. 2001). The evidence base for treatments considered in this guideline is based primarily on randomised controlled trials (RCTs). panic and various personality disorders (Brown et al. in which standardised criteria have been used to determine entry into the trial. In contrast with the previous guideline.

Depression control for or examine these variations. of particular concern are patients with bipolar disorder presenting with depression. Cushing’s syndrome. although it is likely that. 2002). However. 2000). Most people now believe that all of these factors influence a person’s vulnerability to depression. including diabetes. and. The differential diagnosis of depression can be difficult. Personality traits such as ‘neuroticism’ also increase the risk of depression when faced with stressful life events (Fava & Kendler.2 AETIOLOGY The enormous variation in the presentation. 1984). Malhi et al. genetic and family factors. 2000). This serves to emphasise that choice of treatment is a complex process and involves negotiation and discussion with patients. 1978) processes and/or factors. cardiac disease. and physical and sexual abuse almost certainly increase a person’s vulnerability to depression in later life (Fava & Kendler.. adverse childhood experiences. precisely how these factors interact and influence that vulnerability will vary (Harris. Harris. and social (Brown & Harris. 2000. The issue of differential diagnosis in this area is covered in the NICE guideline on bipolar disorder (NICE. psychological (Freud. In the stress-vulnerability model (Nuechterlein & Dawson. such as the end of a relationship (Hammen et al. 1917). neglect. Advances in neuroimaging have reinforced the idea of depression as a disorder of brain structure and function (Drevets et al. but the uncertainty needs to be discussed with the patient and benefits from treatment carefully monitored. endocrine and neurophysiological (Goodwin. marital discord and divorce. 2006c). 2000). given the current limited knowledge about which factors are associated with better antidepressant or psychotherapy response. An emphasis upon physical and especially endocrine theories of causation has been encouraged by the observation that some physical illnesses increase the risk of depression. including genetic (Kendler & Prescott.. hyperthyroidism. 2000). Trials of treatment in unclear cases may be warranted.. 2008) and psychological findings emphasise the importance of cognitive and emotional processes (Beck. course and outcomes of depressive illnesses is reflected in the breadth of theoretical explanations for their aetiology. different personalities have different expectancies of stressful life events and some personalities have different rates of dependent life events that are directly related to their personality. for different people living in different circumstances. Nevertheless. personality factors and social circumstances. 2. most decisions will rely upon clinical judgement and patient preference until there is further research evidence. 2005).. A family history of depressive illness accounts for around 39% of the variance of depression in both sexes (Kendler et al. vulnerability factors interact with social or physical triggers such as stressful life events or physical illness to result in a depressive episode (for example. 25 . 2000). 2008). hypothyroidism. and early life experiences such as a poor parent–child relationship. 1999). the factors identified as likely to increase a person’s vulnerability to depression include gender. biochemical. Addison’s disease and hyperprolactinaemic amenorrhea (Cassano & Fava. 2001).

comorbidity with other psychiatric or physical disorders and certain types of personality disorder). These are likely to relate to. and having no paid employment outside the home (Brown & Harris. depression is projected to become the second leading cause of disability with estimates indicating that unipolar depressive disorders account for 4. These are aspects that self-help interventions and organisations often target.Depression The possession of a specific variation in particular genes has also been reported to make individuals more likely to experience depression when faced with life events (for example.. 2003). the non-specific effects of interventions and the placebo effect (see Section 2.. cannot be doubted even from a brief examination of the epidemiology of depression (see above). triggering a depressive episode. why it varies so much between individuals and why for some it becomes persistent. may lead to secondary disability that compounds.4% of the global disease burden or the equivalent of 65 26 . 1991). However. 1978) or befriending. it is equally important to consider factors that maintain or perpetuate depression because these are potential targets for intervention.3 ECONOMIC COSTS OF DEPRESSION There is now widespread recognition of the significant burden that depression imposes on people and their carers. in which vulnerabilities interact with stressful life events. 1999). such as poverty. The role of current social circumstances in increasing the risk of depression. conversely. unemployment and chronic physical or mental illness. London.3). included: having three or more children under the age of 14 years living at home. 2. such as separation or loss of a loved one. but about which there is little systematic evidence. Features include loss of self-esteem and independence. 1978). do protect against depression following a stressful life event (Harris et al. it is also the case that some factors. health services and communities throughout the world. such as having a supportive and confiding relationship with another person (Brown & Harris. Although many studies have reported on factors that predict outcome (including earlier age of onset. As mentioned previously. The ‘neatness’ of this social model of depression. Caspi et al.4. greater severity and chronicity. the depression itself. feelings of helplessness and hopelessness (which increase the risk of suicide) and loss of engagement in outside activities with social withdrawal. and benefit from. Lack of a confiding relationship appears to be a strong risk factor for depression (Patten. ongoing social stresses. not having a confiding relationship with another person. homelessness. and is difficult to distinguish from. It is also clinically apparent that depression. In the UK. an influential study found that social vulnerability factors for depression in women in Camberwell. In addition to considering the aetiology of the onset of depressive episodes. many such events are not followed by a depressive disorder in those with vulnerabilities. is not always supported by the ‘facts’: some episodes of depression occur in the absence of a stressful event and. there is a lack of understanding about what determines how long a depressive episode lasts. especially when it persists. by 2020.

2002). of which 84% was attributable to antidepressant medication. Furthermore. taken from Psychiatric Morbidity Survey data. Based on UK labour market survey data. The study measured the direct treatment costs of depression. 2001). these figures were projected to be £3 billion and £12. This reduced workplace productivity is unlikely to be 27 . It was estimated that there were 1.45 million by 2026. By 2026. in England for the next 20 years.24 million people with depression in England.5 billion. supported accommodation. A recent review was conducted by the King’s Fund in 2006 to estimate mental health expenditure. and this was projected to rise by 17% to 1.. Within the UK setting. 1997b. the total cost of services for depression in England in 2007 was estimated to be £1. depression places an enormous burden on both the healthcare system and the wider society. WHO. the proportion of lost employment costs (78 to 90%) of the total costs was similar across both studies. Overall. to 2026 (McCrone et al. the indirect costs of depression were estimated to be far greater: total morbidity costs were £8 billion and mortality costs were £562 million. 2002). In comparison with the findings of earlier UK-based cost-of-illness studies. Other studies have also supported these findings. However. One of the key findings from the cost-of-illness literature is that the indirect costs of depression far outweigh the health service costs.2 billion. direct treatment costs shifted from hospital admissions (including specialised psychiatric institutions) towards medication. including primary and secondary care costs as well as indirect costs of lost working days (morbidity) and lost life years (mortality). including prescribed drugs. its role as probably the most important risk factor for suicide (Knapp & Illson. with population estimates for 2007 through to 2026. inpatient care. The direct treatment costs were estimated at £370 million. 2001). in terms of ‘work cut-back days’ (Kessler et al. One UK study estimated the total cost of depression in adults in England in 2000 (Thomas & Morris. Thomas and Morris (2003) suggest that the effect on lost employment and productivity is 23 times larger than the costs falling to the health service. the Psychiatric Morbidity Survey of adults aged 16 to 74 years in 2000 reported a prevalence rate for depression of 26 per 1000 people with slightly higher rates for women compared with men (Singleton et al. Due to its high prevalence and treatment costs. A prevalence-based approach was used by applying rates of depression from Office of National Statistics data to population data for England in 2000. Based on these figures the authors estimated total costs for depression.7 billion.. a US-based study suggests that depression is a major cause of reduced productivity while at work. antidepressant medication accounted for only 1% of total service costs while inpatient and outpatient care accounted for over 50%.Depression million DALYs (Murray & Lopez. The study combined prevalence rates of depression. other NHS services. However. as well as its large impact on workplace productivity.. while lost employment increased this total to £7. social services and lost employment in terms of workplace absenteeism. including depression. Almond and Healey (2003) estimated that respondents with self-reported depression/anxiety were three times more likely to be absent from work (equivalent to 15 days per year) than workers without depression/anxiety. In contrast to the study by Thomas and Morris (2003). reflecting changes in patterns of care over time away from expensive inpatient care to relatively less expensive outpatient-based care. respectively. 2008). 2003).

49 are not recognised as depressed. feeling it was a problem one should be able to cope with (28%). 1998. The stigma associated with depression cannot be ignored in this context (Priest et al. especially in primary care. 2. 1996).. There is considerable variation among individual GPs in their referral rates to mental health services.4 TREATMENT AND MANAGEMENT IN THE NATIONAL HEALTH SERVICE Treatment for depressive illnesses in the NHS is hampered by the unwillingness of many people to seek help for depression and the failure to recognise depression. Other intangible costs of depression include the impact on the quality of life of people with depression and their carers. The improved recognition and treatment of depression in primary care is central to the WHO strategy for mental health (WHO. 1980).4. mainly because most of such patients are consulting for a somatic symptom and do not consider themselves mentally unwell.. only 80 will consult their GP. thinking the problem would get better by itself (15%). most are treated in primary care and about one in four or five are referred to secondary mental health services. and being afraid of the consequences (for example. single women and those below 35 years of age (Goldberg & Huxley. 1992. The most common reasons given for reluctance to contact the family doctor include: not thinking anyone could help (28%). Thompson et al. 2003). Certainly.1 Detection. 2. 10%) (Meltzer et al. it is anticipated that these costs will continue to rise significantly in future years.. people with more severe illnesses. GPs are immensely variable in their ability to recognise depressive illnesses. not thinking it was necessary to contact a doctor (17%). Üstün & Sartorius. hospitalisation.Depression adequately measured by absenteeism rates and further emphasises the ‘hidden costs’ of depression (Knapp. family members. with some recognising virtually all the patients found to be depressed at independent research interview. recognition and referral in primary care Of the 130 cases of depression (including mild cases) per 1000.. feeling too embarrassed to discuss it with anyone (13%). GPs’ communication skills make a vital contribution to determining 28 . and others recognising very few (Goldberg & Huxley. It is therefore important that efficient use of available healthcare resources is made. 2001). tests. but those seen by specialist services are a highly selected group – they are skewed towards those who do not respond to antidepressants. being sectioned. 2001).. 1995). Furthermore. 2000). 1995). Of those that are recognised as depressed. Of the 80 depressed people per 1000 who do consult their GP. treatment. the healthcare system and on the broader economy through lost productivity and workplace absenteeism. This group also has milder illnesses (Goldberg et al. despite the presence of symptoms of depression (Kisely et al. the cost-of-illness calculations presented here show that depression imposes a significant burden on people and their carers. to maximise health benefits for people with depression.

With 50% of people with depression never consulting a doctor. In combination with the strong mediating effect of empathy and psychoeducation they suggest that other. Ormel and colleagues (1990) suggested that the benefits of recognition of common mental disorders could not be attributed entirely to specific mental health treatments. 2000.2 Assessment and co-ordination of care Given the low detection and recognition rates. Kendrick et al. 1993). 1999. lectures and discussion groups have not improved recognition or outcomes (Thompson et al. For example. it is essential that primary care and mental health practitioners have the required skills to assess people with depression. those with more severe disorders.. 1995). Moreover. The effective assessment of a patient.. 1988.. such as acknowledgement of distress. and many more whose depression goes unrecognised and untreated. but did improve outcomes. although similar interventions combined with skills training may improve detection and outcomes in terms of symptoms and level of functioning (Tiemens et al. are especially likely to be recognised as depressed while those presenting with somatic symptoms for which no obvious cause can be found are less likely to be recognised. 2. In addition. about half of the observed improvement in patient outcomes was mediated by the combined improvements in process of care.4. Other factors. who may not themselves wish to complain of their distress unless they are asked directly about it (Goldberg & Bridges... and providing hope and social support. 2004). and those presenting with psychological symptoms. reinterpretation of symptoms. while having face validity.Depression their ability to detect emotional distress and those with superior skills allow their patients to show more evidence of distress during their interviews.. This view has gained confirmation from a Dutch study in which providing skills training for GPs did not improve detection. were suggested to contribute to better patient outcomes. The inference that these health gains are the result of improved detection and better access to specific treatments. 2001). their social circumstances and relationships. with all of the attendant negative personal and social consequences. could be changed. thus making detection easy. the communication skills needed by GPs can be learned and incorporated into routine practice with evident improvement in patient outcomes (Gask et al. 2001). Goldberg et al. Roter et al. this is clearly a problem for primary care. 95% never entering secondary mental health services. probably also nonspecific. and high personal and social costs. Attempts to improve the rate of recognition of depression by GPs using guidelines. has been contested. This is especially important in view of the fact that depression is associated with an increased suicide rate. Ostler et al. The evidence suggests that these very undesirable circumstances. Those GPs with poor communication skills are more likely to collude with their patients. 1988. a strong tendency for recurrence.. in which large numbers of people each year experience depression.. aspects of the process of care must be responsible for the training effect on symptoms and disability (Van Os et al. including risk assessment and 29 . and the risk they may pose to themselves and others. In summary.

those put on waiting lists do improve steadily with time. and non-specific effects. Antidepressants (or other) treatments for depression may offer little or no advantage.. more heterogeneous diagnostic categories.Depression the subsequent co-ordination of their care (through the use of the Care Programme Approach [CPA] in secondary care services). It may not always be possible to achieve remission. with 26% recovering in 1 month. The evidence does support the efficacy of specific treatments with more severe depression 30 . The placebo effect in trials of psychiatric drugs is often so large that specific pharmacological effects can be hard to identify. It should be noted that there is a high relapse rate associated with depression (see Section 2. Studies comparing any treatment with a waiting list control or treatment as usual (TAU) in which there is minimal intervention are therefore difficult to interpret and improvements could simply be due to the increased support. and 50% improved within 6 months. especially with regard to drug company funded trials (Lexchin et al. with 50% remission in 25 weeks. therefore. There can also be suspicion of publication bias. Although there is insufficient space here to allow proper discussion. Melander et al. remission. Despite their greater severity and other differences. 85% in 1 year. Furukawa and colleagues (2000) showed that patients treated by psychiatrists with antidepressants showed greater improvements than untreated patients: the median time to recovery was 3 months. including the severity of depression. The relative importance of these depends on many factors. but it is usually possible to improve symptoms and functioning to an important degree. change in symptoms and relapse. on average. Among those seeking treament for depression. who often improve spontaneously or who respond well to non-specific measures such as support and monitoring. especially when given to people who fall into one of the larger. it should be noted that non-specific/placebo effects apply not only to treatment with medication but also to other treatments. Posternak and Miller (2001) studied 221 patients assigned to waiting lists in 19 treatment trials of specific interventions and found that 20% improved within 4 to 8 weeks. to prevent relapse. of treatment and the placebo The aim of intervention is to restore health through the relief of symptoms and restoration of function and. For this reason the GDG examined a range of outcomes (where available). and 88% in 2 years. is highly likely to improve outcomes and should. 63% in 6 months. An earlier study by Coryell and colleagues (1994) followed up 114 patients with untreated depression for 6 months: the mean duration of an episode was 6 months. the key goal of an intervention should be complete relief of symptoms (remission).4.1. the degree of impairment to everyday functioning experienced and the patient’s psychiatric history. above). 2003). 2.. engagement and monitoring that the intervention involves. 2005). over placebo for patients with subthreshold depressive symptoms or mild depression. in the longer term. be comprehensive. Where possible. including response.3 Aim. 2003. They estimated that 60% of responders to placebo and 30% of responders to antidepressants may experience spontaneous resolution of symptoms (if untreated).2. which is associated with better functioning and a lower likelihood of relapse (Kennedy & Foy.

there is preliminary evidence that pharmacogenetic variations may affect the efficacy and tolerability of antidepressant drugs. some 5HT and others affecting both to varying degrees and in different ways. meaning and psychological treatment of 31 . Freud published ‘Mourning and melancholia’. 2. 2. which is probably the first modern psychological theory on the causes.4. the side effects resulting from their ability to influence anticholinergic. although their mood-elevating effects are again thought to be mediated through increasing intra-synaptic levels of monoamines. 5hydroxytryptymine (5HT) and dopamine (DA). although the use of these drugs is usually reserved for people with severe. In response to the side-effect profile and the toxicity of TCAs in overdose. given the lack of specific treatments for people with depression. Although the introduction of the TCAs was welcome. The antidepressant properties of monoamine-oxidase inhibitors (MAOIs) were discovered by chance in the 1950s. including: selective serotonin reuptake inhibitors (SSRIs). 1982). such as trazodone. Tricyclic antidepressants (TCAs) were introduced in the 1950s.5 Psychological treatments In 1917. the TCAs predominantly affect the reuptake of NA and 5HT rather than DA (Mindham. overdose with TCAs (with the exception of lofepramine) carries a high mortality and morbidity. psychotic or chronic depressions. In addition. and a range of other chemically unrelated antidepressants. which is particularly problematic in the treatment of people with suicidal intentions. 2009). thought to be responsible for their mood-elevating properties. some primarily affecting NA. In fact. the first being imipramine (Kuhn. The mode of action of this class of drug. 2009) and antipsychotics (BNF 57.Depression and in those with depression that persists over time. A full review of the evidence base for the use of the different types of antidepressants is presented in Chapter 10. in parallel with TCAs. Other drugs used either alone or in combination with antidepressants include lithium salts (BNF 57. Their effects and side effects vary considerably. However at present it is not possible to clearly identify people with depression who will respond to the specific aspects of a treatment as opposed to the non-specific effects associated with having a treatment. histaminergic and other receptor systems reduced their acceptability. such as fluoxetine. is their ability to block the synaptic reuptake of monoamines. It is likely that future research on this topic will lead to the development of clinically meaningful pharmacogenetic markers. including mirtazapine (BNF 57. 2009). new classes of antidepressants have been developed.4. or as prophylactics. including noradrenaline (NA). 1958). Moreover. drugs chemically related to but different from the TCAs.4 Pharmacological treatments The mainstay of the pharmacological treatment of depression for the last 40 or more years has been antidepressants. but at the moment the data is insufficient to make recommendations.

see Department of Health. but address the difficulties of treating people with depression using a less rigid psychoanalytic approach (Fonagy. a ‘stepped-care’ model is developed that draws attention to the different needs that depressed individuals have – depending on the characteristics of their depression and their personal and social circumstances – and the responses that are required from services. the emergence of cognitive and behavioural approaches to the treatment of mental health problems has led to a greater focus upon the evidence base and the development of psychological treatments specifically adapted for people with depression (for example. community teams. 1996). interpersonal therapy (IPT). Since that time. Many. behavioural activation. choosing the right service configuration for the delivery of care to specific groups of people has gained increasing interest with regard to both policy (for example.. In any event. 1999). guided self-help.7 Stepped care In Figure 1. day hospital care. problem-solving therapy. numerous theories and methods for the psychological treatment of psychological disorders have been elaborated and championed. and couples therapy. 2. for example. in particular primary care counselling.4. computerised cognitive behavioural therapy (CCBT) and screening. Service-level interventions considered for review in this guideline include: organisational developments. Marshall et al.. In the last 15 years in the UK there has been a very significant expansion of psychological treatments in primary care for depression. and the variation in the way care is delivered (inpatient. but not all.. Stepped care provides a framework in which to organise the 32 . and research into their efficacy is more recent still (Roth & Fonagy. counselling. and so on). short-term psychodynamic psychotherapy. non-statutory support and other social supports. evaluating day hospital treatment. 1979). Research using RCT designs has a number of difficulties. such therapies are derived from Freudian psychoanalysis. although psychological treatments specifically for depression were developed only over the last 30 to 40 years. Psychological treatments for depression currently claiming efficacy in the treatment of people with depressive illnesses and reviewed for this guideline in Chapter 8 include: cognitive behavioural therapies. see Beck et al. Psychological treatments have expanded rapidly in recent years and generally have more widespread acceptance from patients (Priest et al.Depression depression. crisis teams. 2001).4. day hospital. 1996). 2003).6 Service-level and other interventions Given the complexity of healthcare organisations. 2. Other types of interventions reviewed for this guideline include: physical activity programmes. using comparators such as ‘standard care’ in the US make the results difficult to generalise or apply to countries with very different types of ‘standard care’. outpatient. and research (for example.

Treatment of depression in primary care. high-intensity psychological interventions. ECT. there are a few patients who will need admission to an inpatient psychiatric bed. which can be accompanied by general advice on such matters as restoring natural sleep rhythms and getting more structure into the day. moderate and severe depression STEP 2: Persistent subthreshold depressive symptoms. However. high-intensity psychological interventions. provision of services supporting patients. the main indications being failure of the depression to respond to treatment offered in primary care. and/or is associated with significant psychiatric comorbidity or psychosocial factors. 2009c). and these patients are therefore offered more complex interventions.. psychological or social interventions. Various interventions are effective. active monitoring and referral for further assessment and interventions 1 Complex depression includes depression that shows an inadequate response to multiple treatments. combined treatments. collaborative care2 and referral for further assessment and interventions Low-intensity psychosocial interventions. medication and referral for further assessment and interventions STEP 1: All known and suspected presentations of depression Assessment. psychological interventions. support. many GPs prefer an ‘active monitoring’ approach. Those patients who are actively suicidal or whose depression has psychotic features will need specialist referral. they can receive 24-hour care and various special interventions. severe self-neglect STEP 3: Persistent subthreshold depressive symptoms or mild to moderate depression with inadequate response to initial interventions. combined treatments. risk to life. 2 Only for depression where the person also has a chronic physical health problem and associated functional impairment (see NICE. is complicated by psychotic symptoms. psychoeducation. some prefer to avoid medical interventions and others will improve in any case without them. in depression of only mild severity. other people prefer to accept. 1995). or indeed require. Of those people whom primary healthcare professionals recognise as having depression. incomplete response or frequent recurrences of depression. only about one in five of the patients at this level will need referral to a mental healthcare professional. carers and healthcare professionals in identifying and accessing the most effective interventions.Depression Figure 1: The stepped-care model Focus of the intervention STEP 4: Severe and complex1 depression. Thus. medical. often falls short of optimal guideline recommended practice (Donoghue & Tylee. multiprofessional and inpatient care Medication. mild to moderate depression Nature of the intervention Medication. Finally. crisis service. however. As we have seen. Here. 33 . delivered by a range of workers in primary care. 1996) and outcomes are correspondingly below what is possible (Rost et al.

which sets the parameters of the update and provides a focus and steer for the development work ● update the clinical questions developed for the previous guideline ● develop criteria for updating the literature search and conduct the search ● design validated protocols for systematic review and apply to evidence recovered by search ● synthesise and (meta-) analyse data retrieved. the concerns of people with depression and their carers regarding health and social care have been highlighted and addressed by recommendations agreed by the whole GDG. with support from the NCCMH staff. 3.1 OVERVIEW The update of this guideline drew upon methods outlined by NICE (The Guidelines Manual. NCCMH. 2004a. and produce evidence summaries (for both the clinical and health economic evidence) ● decide if there is sufficient new evidence to change existing recommendations and develop new recommendations where necessary. 2003). 2004). The purpose of the scope is to: ● provide an overview of what the guideline will include and exclude ● identify the key aspects of care that must be included 2The methodology for the previous guideline can be found in Appendix 21. The update will provide recommendations for good practice that are based on the best available evidence of clinical and cost effectiveness. NICE. A team of healthcare professionals. 2007c). guided by the clinical questions.2 THE SCOPE NICE commissioned the NCCMH to review recent evidence on the management of depression and to update the existing guideline Depression: Treatment and Management of Depression in Primary and Secondary Care (NICE. to ensure a service user and carer focus. There are six basic steps in the process of updating a guideline: ● define the scope. In addition. lay representatives and technical experts known as the Guideline Development Group (GDG). which had been incorporated into the previous guideline. evidence-based guideline. The scope for the update also included updating the NICE technology appraisal on the use of ECT (NICE. 34 .Methods used to develop this guideline 3 METHODS USED TO DEVELOP THIS GUIDELINE2 3. undertook the update of a patient-centred. The NCCMH developed a scope for the guideline update (see Appendix 1).

These groups were designed to efficiently manage the large volume of evidence needing to be appraised prior to presenting it to 35 . 3. and GDG members formed smaller topic groups to undertake guideline work in that area of clinical practice. psychiatric pharmacy. in a plenary session. and recommendations formulated. 2007c).2 Topic groups The GDG divided its workload along clinically relevant lines to simplify the guideline development process.uk). and the concerns of people with depression and carers were routinely discussed as part of a standing agenda item. The guideline development process was supported by staff from the NCCMH.3. The NCCMH and NICE reviewed the scope in light of comments received and the revised scope was signed off by the GRP. Comments were invited from stakeholder organisations and the Guideline Review Panel (GRP). Further information about the GRP can also be found on the NICE website. 3.org.3. The draft scope was subject to consultation with registered stakeholders over a 4-week period.1 Guideline Development Group meetings Fourteen GDG meetings were held between November 2007 and January 2009. managed the process and contributed to drafting the guideline. academic experts in psychiatry and psychology. all GDG members declared any potential conflicts of interest. nursing and general practice. The GDG was recruited according to the specifications set out in the scope and in line with the process set out in the NICE guideline manual (NICE. During each day-long GDG meeting. (2) psychological and psychosocial interventions and (3) services. who undertook the clinical and health economics literature searches.3 THE GUIDELINE DEVELOPMENT GROUP The GDG consisted of: professionals in psychiatry. and people with depression and a carer. the scope was posted on the NICE website (www. Three topic groups were formed to cover: (1) pharmacological and physical interventions. reviewed and presented the evidence to the GDG. During the consultation period. clinical psychology. 3. clinical questions and clinical and economic evidence were reviewed and assessed.nice.Methods used to develop this guideline ● set the boundaries of the development work and provide a clear framework to enable work to stay within the priorities agreed by NICE and the NCC and the remit from the Department of Health/Welsh Assembly Government ● inform the development of updated clinical questions and search strategy ● inform professionals and the public about the expected content of the guideline ● keep the guideline to a reasonable size to ensure that its development can be carried out within the allocated period. At each meeting.

Topic group leaders reported the status of the group’s work as part of the standing agenda. In drafting the guideline. one of whom was also a carer. reviewed and prepared the evidence with the systematic reviewer before presenting it to the GDG as a whole and helped the GDG to identify further expertise in the topic. and trial data if the GDG could be provided with full access to the complete trial report. where necessary.3. they contributed to writing the guideline’s introduction and Chapter 4 and identified recommendations from the service user and carer perspective.3 People with depression and carers Individuals with direct experience of services gave an integral service-user focus to the GDG and the guideline. They contributed as full GDG members to writing the clinical questions. and bringing service-user research to the attention of the GDG. The service user and carer representatives jointly ran the group and presented their findings at GDG meetings. or provided expertise in methodological aspects of evidence synthesis. A group was also convened comprising the service user and carer representatives and members of the NCCMH review team to develop the chapter on experience of care (Chapter 4).3.Methods used to develop this guideline the GDG as a whole. Each topic group was chaired by a GDG member with expert knowledge of the topic area (one of the healthcare professionals).4 Special advisers Special advisers. helping to ensure that the evidence addressed their views and preferences. who had specific expertise in one or more aspects of treatment and management relevant to the guideline. commenting on specific aspects of the developing guideline and. They also introduced and led the GDG discussion of the evidence review for that topic and assisted the GDG Chair in drafting the section of the guideline relevant to the work of each topic group. They informed the group about completed trials at the pre-publication stage. assisted the GDG. Appendix 6 lists the researchers who were contacted.5 National and international experts National and international experts in the area under review were identified through the literature search and through the experience of the GDG members. systematic reviews in the process of being published.3. 36 . Topic groups refined the clinical questions and the clinical definitions of treatment interventions. 3. These experts were contacted to recommend unpublished or soon-to-be published studies to ensure that up-to-date evidence was included in the development of the guideline. 3. Appendix 3 lists those who agreed to act as special advisers. making presentations to the GDG. The GDG included three people with depression. studies relating to the cost effectiveness of treatment. 3. highlighting sensitive issues and terminology relevant to the guideline.

over and above its general significance in relation to specific interventions. The draft clinical questions were discussed by the GDG at the first few meetings and amended as necessary. rates of relapse. intervention. For questions about interventions. Where appropriate. questions related to issues of service delivery are occasionally specified in the remit from the Department of Health/Welsh Assembly Government. appropriate clinical questions were developed to be clear and concise. Questions submitted by stakeholders were also discussed by the GDG and included where appropriate. In these cases. late morbidity and readmission.4 CLINICAL QUESTIONS Clinical questions were used to guide the identification and interrogation of the evidence base relevant to the topic of the guideline. Areas where this is particularly likely to occur relate to assessment of risk. comparison and outcome) framework was used. The review focused on providing evidence to answer the primary questions. the comparisons (other main treatment options) and the outcomes (the measures of how effective the interventions have been) (see Table 2). mortality. the interventions (what is being done). physical and social functioning and other measures. Table 2: Features of a well-formulated question on effectiveness intervention – the PICO guide Patients/population Which patients or population of patients are we interested in? How can they be best described? Are there subgroups that need to be considered? Which intervention. treatment or approach should be used? What is/are the main alternative/s to compare with the intervention? What is really important for the patient? Which outcomes should be considered: intermediate or short-term measures. subquestions were generated. the prognosis of a particular condition is of fundamental importance. costs? 37 Intervention Comparison Outcome . morbidity and treatment complications.Methods used to develop this guideline 3. return to work. In addition. the PICO (patient. where necessary. For the purposes of the systematic review of clinical evidence. such as quality of life. The final list of clinical questions can be found in Appendix 7. general health status. the questions were refined once the evidence had been searched and. In some situations. for example in terms of early intervention. This structured approach divides each question into four components: the patients (the population under study). the questions were categorised as primary or secondary.

interrupted time-series Accuracy of information (for example. 3. The NCCMH developed this process based on methods set out in The 38 . These are listed in Table 3. Thus.Methods used to develop this guideline Table 3: Best study design to answer each type of question Type of question Effectiveness or other impact of an intervention Best primary study design RCT. other studies that may be considered in the absence of an RCT are the following: internally/externally controlled before and after trial. if evidence is not available. prediction rule) valid gold standard in a randomised trial or inception cohort study Rates (of disease. test.11) and the need for future research is specified. informal consensus methods are used (see Section 3. where ‘best’ is interpreted as ‘least likely to give misleading answers to the question’. cross-sectional study Naturalistic prospective cost study To help facilitate the literature review. registry. There are four main types of clinical question of relevance to NICE guidelines. Comparing the information against a risk factor.5 SYSTEMATIC CLINICAL LITERATURE REVIEW The aim of the clinical literature review was to systematically identify and synthesise relevant evidence from the literature (updating the existing evidence-base where appropriate) to answer the specific clinical questions developed by the GDG. a note was made of the best study design type to answer each question. For each type of question the best primary study design varies.5.5. rare side effects) Costs Cohort. Deciding on the best design type to answer a specific clinical question does not mean that studies of different design types addressing the same question were discarded. patient experience. 3. clinical practice recommendations are evidence-based where possible and. in all cases a well-conducted systematic review of the appropriate type of study is likely to always yield a better answer than a single study. However.1 Methodology A step-wise hierarchical approach was taken to locating and presenting evidence to the GDG.

The initial approach taken to locating primary-level studies depended on the type of clinical question and potential availability of evidence. These included: ● Clinical Policy and Practice Program of the New South Wales Department of Health (Australia) ● Clinical Evidence online ● The Cochrane Collaboration ● New Zealand Guidelines Group ● NHS Centre for Reviews and Dissemination ● Oxford Centre for Evidence-Based Medicine ● Oxford Systematic Review Development Programme ● Scottish Intercollegiate Guidelines Network (SIGN) ● United States Agency for Healthcare Research and Quality. and any additional assessments. a brief descriptive review was initially undertaken by a member of the GDG.Methods used to develop this guideline Guidelines Manual (NICE. searches were for the appropriate study design (see above). the evidence was likely to change a recommendation). 2007c) and after considering recommendations from a range of other sources. Although there are a number of difficulties with the use of RCTs in the evaluation of interventions in mental health. For questions that were unlikely to have a good evidence base. For other clinical questions.2 The review process During the development of the scope. After this point. a more extensive search was undertaken for systematic reviews and guidelines published since the previous depression guideline. 3. a decision was made about which questions were best addressed by good practice based on expert opinion. 39 . Based on the previous guideline and GDG knowledge of the literature. 3. the RCT remains the most important method for establishing treatment efficacy. studies were included only if they were judged by the GDG to be exceptional (for example. the initial evidence base (or updated evidence base) was formed from well-conducted RCTs that addressed at least one of the clinical questions.5.3 The search process for questions concerning interventions For questions related to interventions. Recommendations based on good practice were developed by informal consensus of the GDG. Searches for evidence were updated between 6 and 8 weeks before the guideline consultation.5. which questions were likely to have a good evidence base and which questions were likely to have little or no directly relevant evidence. the criteria for assessing the eligibility of studies. the review process depended on the type of key question (see below). For questions with a good evidence base. Review protocols included the relevant clinical question(s). the search strategy. These were used to inform the development of review protocols for each topic group.

Methods used to develop this guideline The search was exhaustive. For RCTs the search consisted of terms relating to the clinical condition (that is. in some circumstances existing datasets were utilised. In addition. New RCTs meeting inclusion criteria set by the GDG were incorporated into the existing reviews and fresh analyses performed. In addition. MEDLINE and PsycINFO) were used for the initial search for all studies potentially relevant to the guideline update. Where this was the case. The GDG did not limit the search to any particular therapeutic modality. For questions without good-quality evidence (after the initial search). results were scanned liberally to exclude irrelevant papers. the review team used filters developed for systematic reviews. 3Unpublished full trial reports were also accepted where sufficient information was available to judge eligibility and quality (see Section 3. searches were made of the reference lists of all eligible systematic reviews and included studies. thereby yielding the largest number of relevant papers that might otherwise be missed by more specific searches. However. the review team used a purpose-built ‘study information’ database to manage both the included and the excluded studies (eligibility criteria were developed after consultation with the GDG).4 Search filters Search filters developed by the review team consisted of a combination of subject heading and free-text phrases. where necessary. depression) and study design only. CINAHL. for each clinical question. Known experts in the field. 40 . Double checking of all excluded studies was not done routinely. RCTs and other appropriate research designs (Appendix 8). In addition. AMED. formed around additional elements of the question. recent high-quality English-language systematic reviews were used primarily as a source of RCTs (see Appendix 10 for quality criteria used to assess systematic reviews). including interventions and the outcomes of interest.5. were sent letters requesting relevant studies that were in the process of being published (see Appendix 6)3. using several databases and other sources. the tables of contents of appropriate journals were periodically checked for relevant studies. ERIC. Where the evidence base was large. but a selection of abstracts was checked to ensure reliability of the sifting.5.5. EMBASE. (b) conduct a new search for lower levels of evidence or (c) adopt a consensus process (see Section 3. data were cross-checked for accuracy before use.11). OpenSIGLE or Sociological Abstracts). based both on the references identified in early steps and on advice from GDG members.6). 3. Cochrane Library. Standard mental health related bibliographic databases (that is. a decision was made by the GDG about whether to (a) repeat the search using subject-specific databases (for example. Specific filters were developed for the guideline topic and. After the initial search.

Eligible systematic reviews and primarylevel studies were critically appraised for methodological quality (see Appendix 10 for the quality checklists and Appendix 17 for characteristics of each study including quality assessment).5. it was necessary to prioritise the evidence with respect to the UK context (that is.Methods used to develop this guideline 3. However.5 Study selection All primary-level studies included after the first scan of citations were acquired in full and re-evaluated for eligibility (based on the relevant review protocol) at the time they were being entered into the study database. gender. age and ethnicity) ● provider factors (for example. the topic groups took into account the following factors when assessing the evidence: ● participant factors (for example. 2003) or Review Manager 5 (Cochrane Collaboration.5.5.2. the GDG distinguished between outcomes that they considered critical and ones that were important but not critical for the purposes of updating the guideline.7 Data extraction Outcome data were extracted from all eligible studies. It was the responsibility of each topic group to decide which prioritisation factors were relevant to each clinical question in light of the UK context and then decide how they should modify their recommendations.6 Unpublished evidence The GDG used a number of criteria when deciding whether or not to accept unpublished data. which met the minimum quality criteria. The eligibility of each study was confirmed by consensus during topic group meetings. second. where evidence was submitted directly to the GDG. First. 2008). 3. Only critical outcomes were initially extracted for data analysis (further details about the critical outcomes can be found in the evidence chapters). differences in standard care and differences in the welfare system). using Review Manager 4. 41 . the conditions under which the intervention was performed and the availability of experienced staff to undertake the procedure) ● cultural factors (for example. the evidence must have been accompanied by a trial report containing sufficient detail to properly assess the quality of the research. model fidelity. For each major area reviewed. it must have been done so with the understanding that details would be published in the full guideline. external validity). For some clinical questions. 3. the GDG recognised that unpublished evidence submitted by investigators might later be retracted by those investigators if the inclusion of such data would jeopardise publication of their research.10 (Cochrane Collaboration. To make this process explicit.

Where consensus could not be reached. the authors.8 Synthesising the evidence Where possible. tables and appendices of this guideline. early withdrawals were included in both the numerator and denominator. the data were excluded from the analysis (except for the outcome ‘leaving the study early’. re-analyses of the data or sub-analyses were used to answer clinical questions not addressed in the original studies or reviews. A ‘relative risk’ (also called a ‘risk ratio’) is the ratio of the treatment event rate to the control event rate. 1997). Disagreements were resolved with discussion. dichotomous efficacy outcomes were calculated on an intention-to-treat basis (that is. 3. JONES2005). If necessary. a ‘oncerandomised-always-analyse’ basis). Study IDs are composed of the first author’s surname followed by the year of publication. In Figure 2.1. studies that were found and included in this guideline update only are labelled in capital letters (for example. References to included and excluded studies can be found in Appendix 17. Data from studies included in existing systematic reviews were extracted independently by one reviewer and cross-checked with the existing dataset. confidence intervals (CIs) or p-values according to standard formulae (see the Cochrane Handbook for Systematic Reviews of Interventions.Methods used to develop this guideline In most circumstances.0. Where possible. Version 5. the 50% rule was not applied. Studies that were included in the previous guideline (NCCMH. in which case the denominator was the number randomised). standard deviations were calculated from standard errors (SEs). Adverse events were entered into Review Manager as reported by the study authors because it was usually not possible to determine whether early withdrawals had an unfavourable outcome. blind to the journal from which the article comes. An RR of 1 indicates no difference between treatment and control.73 42 . 2004) have a study ID in title case (for example. Where there was limited data for a particular review. the evidence was downgraded due to the risk of bias. Where necessary. Higgins & Green. Masked assessment (that is. Data were summarised using the generic inverse variance method using Review Manager. for a given outcome (continuous and dichotomous) where more than 50% of the number randomised to any group were lost to follow up. see Figure 2). the institution and the magnitude of the effect) was not used since it is unclear that doing so reduces bias (Jadad et al. Smith1999). Where possible. Studies have been given a ‘study ID’ to make them easier to identify in the text. a third reviewer or GDG members resolved the disagreement. Consultation with another reviewer or members of the GDG was used to overcome difficulties with coding. 1996. Dichotomous outcomes were analysed as relative risks (RR) with the associated 95% CI (for an example. Where there was good evidence that those participants who ceased to engage in the study were likely to have an unfavourable outcome. data extracted by one reviewer was checked by a second reviewer. 2008). Berlin. In these circumstances. meta-analysis was used to synthesise the evidence using Review Manager.5. the overall RR of 0..

df = 2 (P = 0.19).3% Test for overall effect: Z = 3.00 -0. intention-to-treat data. [0.40(24.10) 11 Wolf1992 109 91 Subtotal (95% CI) Test for heterogeneity: Chi² = 6.90 100.10] 1.81] 0. [-1.83.28 13.25(1. df = 4 (P = 0.04.20. were preferred over data from completers.50 -0.5 1 2 5 38.56.8% Test for overall effect: Z = 4.09] 0.00) 14 Lee1986 28 44. 0. in other words.21) 10.91 17.40) 20 Freeman1988 20 1.74 [-1.43] -0. Continuous outcomes were analysed as weighted mean differences (WMD).45) 22 Griffiths1994 14 3.84 0.37 (P = 0.26] -0. see Figure 3).92 100.24). control 32 1.59 0.50) 61.83 15.14.13.2 0.79 22.60) 4.98 (P < 0.45] –4 –2 0 2 4 Favours intervention Favours control 43 . [0. the RR reduction is 27%.30 38.0007) 0. [-2.70(4.00001) 3.24.36 -0. [0. using a method such as ‘last observation carried forward’. The I2 statistic describes the proportion of total variation in study estimates that is due to heterogeneity (Higgins & Thompson.25. If provided. I² = 34.10(17.14(2.79 0. both the I2 test of heterogeneity and a visual inspection of the forest plots were used.61.88] Favours intervention Favours control indicates that the event rate (that is.23(27. The I2 statistic was interpreted in the following way: ● 50%: notable heterogeneity (an attempt was made to explain the variation by conducting sub-analyses to examine potential moderators.97) 7. the effect is statistically significant.08 27. or as a standardised mean difference (SMD) when different measures were used in different studies to estimate the same underlying effect (for an example.68 -1. studies with effect sizes greater than two standard deviations from the mean of the Figure 3: Example of a forest plot displaying continuous data Review: NCCMH clinical guideline review (Example) Comparison: 01 Intervention A compared to a control group Outcome: 03 Mean frequency (endpoint) Study or sub-category Intervention A Mean (SD) Control Mean (SD) SMD (fixed) 95% CI Weight % SMD (fixed) 95% CI N N 01 Intervention A vs.30(3.66. -0.00 0.09] 0. I² = 29.65 -0. To check for consistency between studies. If the CI does not cross the ‘line of no effect’. control 13/23 27/28 Griffiths1994 11/15 14/15 Lee1986 21/28 24/27 Treasure1994 45/66 65/70 Subtotal (95% CI) Test for heterogeneity: Chi² = 2.49 -0. The CI shows with 95% certainty the range within which the true treatment effect should lie and can be used to determine statistical significance.21. [-1. [-1.30(5.10] -0. non-remission rate) associated with intervention A is about three quarters of that with the control intervention or.04) 24 Treasure1994 15 5.Methods used to develop this guideline Figure 2: Example of a forest plot displaying dichotomous data Review: NCCMH clinical guideline review (Example) Comparison: 01 Intervention A compared to a control group Outcome: 01 Number of people who did not show remission Study or sub-category Intervention A n/N Control n/N RR (fixed) 95% CI Weight % RR (fixed) 95% CI 01 Intervention A vs.84] 1.41.10(4. In addition.60) 25. 2002). [-1.70(3.73 [0.

observational study. If studies with heterogeneous results were found to be comparable with regard to study and participant characteristics.8 for how consistency was measured) ● indirectness (that is. Each evidence profile also included a summary of the findings: the number of patients included in each group. 30 to 50%: moderate heterogeneity (both the chi-squared test of heterogeneity and a visual inspection of the forest plot were used to decide between a fixed and random-effects model). both the quality of the evidence and the results of the evidence synthesis (see Table 4 for an example of an evidence profile). an estimate of the magnitude of the effect and the overall quality of the evidence for each outcome. directness and any other considerations) and graded using the following definitions: ● High further research is very unlikely to change our confidence in the estimate of the effect ● Moderate further research is likely to have an important impact on our confidence in the estimate of the effect and may change the estimate 44 . In the random-effects analysis. 1986]. For observational studies. where appropriate. limitations to study quality. if plausible confounding would have changed the effect. interventions and participants match those of interest) ● imprecision (based on the CI around the effect size). ● ● 3.5. or any other evidence) ● limitations (based on the quality of individual studies. For each outcome. see Appendix 10 for the quality checklists) ● inconsistency (see Section 3. with the exception of diagnostic studies (methods for these studies are at present not sufficiently developed). the quality may be increased if there is a large effect. how closely the outcome measures.Methods used to develop this guideline remaining studies were excluded using sensitivity analyses. or if there is evidence of a dose–response gradient (details would be provided under the other considerations column). The quality of the evidence was based on the quality assessment components (study design. consistency. quality may be reduced depending on the following factors: ● study design (randomised trial.9 Presenting the data to the GDG Study characteristics tables and. forest plots generated with Review Manager were presented to the GDG to prepare a GRADE evidence profile table for each review and to develop recommendations. 30%: mild heterogeneity (a fixed-effects model was used to synthesise the results). Evidence profile tables A GRADE evidence profile was used to summarise. heterogeneity is accounted for both in the width of CIs and in the estimate of the treatment effect. With decreasing heterogeneity the random-effects approach moves asymptotically towards a fixed-effects model). a random-effects model was used to summarise the results [DerSimonian & Laird.5.

6 to 0.Table 4: Example of GRADE evidence profile Summary of findings No.13 (−0.94)3 18 fewer per 100 (from 2 fewer to 25 fewer)  MODERATE 6 Randomised trial No serious limitations No serious inconsistency Outcome 3 No serious indirectness Serious2 None 83 81 – MD −1.03)  MODERATE 3 Randomised trial No serious limitations No serious inconsistency Outcome 5 No serious indirectness Serious4 None 109 114 – SMD −0. would represent a benefit that. of patients Indirectness Imprecision Other considerations Intervention Control Relative (95% CI) Absolute Effect Quality Quality assessment No.44 (0. one would not recommend the intervention.81 to 0. would still be worth it.8)  MODERATE 3 Randomised trial No serious limitations No serious inconsistency Outcome 4 No serious indirectness Serious4 None 88 93 – SMD −0. given the downsides of the intervention.94 (0.51 (−3.34)  MODERATE 4 Randomised trial No serious limitations No serious inconsistency 1The Methods used to develop this guideline 2The upper confidence limit includes an effect that.21 to 0. given the downsides. if it were real.39 to 2. lower confidence limit crosses a threshold below which. 45 .56 to 0. 495% CI crosses the minimal importance difference threshold.23) 0 fewer per 100 (from 3 fewer to 6 more)  MODERATE 6 Randomised trial No serious limitations No serious inconsistency Outcome 2 No serious indirectness Serious2 None 55/236 63/196 RR 0.26 (−0. of studies Design Limitations Inconsistency Outcome 1 No serious indirectness Serious1 None 8/191 7/150 RR 0. 3Random-effects model.

Methods used to develop this guideline ● Low further research is very likely to have an important impact on our confidence in the estimate of the effect and is likely to change the estimate ● Very low any estimate of effect is very uncertain. or where the GDG was of the opinion (on the basis of previous searches or their knowledge of the literature) that there were unlikely to be such evidence. 3. the values of the GDG and society. This process focused on those questions that the GDG considered a priority. These included economic considerations. the associated recommendations were drafted. summary tables incorporating important information from the GRADE profiles were developed (these tables are presented in the evidence chapters). taking into account the trade-off between the benefits and downsides of treatment as well as other important factors.5. ● Evidence from the existing review or new review was then presented in narrative form to the GDG and further comments were sought about the evidence and its perceived relevance to the clinical question. Where this was not possible.10 Forming the clinical summaries and recommendations Once the GRADE profile tables relating to a particular clinical question were completed. high-quality research.11 Method used to answer a clinical question in the absence of appropriately designed. The process involved a number of steps: ● A description of what is known about the issues concerning the clinical question was written by one of the topic group members. a brief review of the recent literature was initiated. The confidence surrounding the evidence in the depression guideline also influenced the GDG’s decision to extrapolate.. For further information about the process and the rationale of producing an evidence profile table. with help from the systematic reviewer. 1998). 3. a narrative review that most directly addressed the clinical question. see GRADE (2004).5. This existing narrative review or new review was used as a basis for beginning an iterative process to identify lower levels of evidence relevant to the clinical question and to lead to written statements for the guideline. Informal consensus The starting point for the process of informal consensus was that a member of the topic group identified. and the GDG’s awareness of practical issues (Eccles et al. Once the GRADE profiles and clinical summaries were finalised and agreed by the GDG and the evidence from depression in the general populations was taken into account. The systematic reviewer in conjunction with the topic group lead produced a clinical evidence summary. high-quality research In the absence of appropriately designed. 46 . either an informal or formal consensus process was adopted.

Methods employed 47 . where economic evidence was lacking or was considered inadequate to inform decisions. At this time. the following economic issues were identified by the GDG in collaboration with the health economist as key priorities for further economic analysis (either costing of interventions or full economic modelling) in the guideline update: ● a cost analysis of low-intensity psychological interventions ● cost-utility of pharmacological interventions ● cost-utility of pharmacological therapy versus combined psychological and pharmacological therapy. a full systematic review was done. Moreover. the report was then sent to appointed experts outside the GDG for peer review and comment. In addition to the systematic review of economic literature. After this final stage of comment. The rest of this section describes the methods adopted in the systematic literature review of economic studies undertaken for this guideline update. Recommendations were then developed and could also be sent for further external peer review. a significant body of primary-level studies (of appropriate design to answer the question) were identified. additional information was sought and added to the information collected. Following this. a series of statements that directly addressed the clinical question were developed. the statements and recommendations were again reviewed and agreed upon by the GDG. This was achieved by: ● a systematic literature review of existing economic evidence ● economic modelling. during the course of preparing the report.6 HEALTH ECONOMICS METHODS The aim of health economics was to contribute to the guideline’s development by providing evidence on the cost effectiveness of interventions for people with depression covered in the guideline. The information from this process was then fed back to the GDG for further discussion of the statements. areas for further economic analysis were prioritised based on anticipated resource implications of the respective recommendations as well as on the quality and availability of respective clinical data. 3. subject possibly to further reviews of the evidence. These topics were selected after considering potential resource implications of the respective recommendations. If. This may have included studies that did not directly address the clinical question but were thought to contain relevant data. literature on health-related quality of life of people with depression was systematically searched to identify studies reporting appropriate utility weights that could be utilised in a cost-utility analysis. on occasions and as deemed appropriate by the GDG. Systematic search of the economic literature was undertaken on all areas that were updated since the previous guideline.Methods used to develop this guideline ● ● ● ● ● ● Based on the feedback from the GDG.

as well as in the HTA database. The quality assessment was based on the checklists used by the British Medical Journal to assist referees in appraising full and partial economic analyses (Drummond & Jefferson. For these databases. MEDLINE. were duplicates. database-specific strategy. The searches were updated regularly. ● Studies published from 1998 onwards were included. 3. Initial searches were performed in November 2007. For the HTA and NHS EED databases. a health economics search filter adapted from the Centre for Reviews and Dissemination at the University of York was used in combination with a general search strategy for depression. Publications that were clearly not relevant were excluded (stage 2). Additional searches were performed in specific health economics databases (NHS Economic Evaluation Database [EED].6. Office of Health Economics Health Economic Evaluations Database [OHE HEED]).Methods used to develop this guideline in de novo economic modelling carried out for this guideline update are described in the respective sections of the guideline.1 Search strategy For the systematic review of economic evidence the standard mental-health-related bibliographic databases (EMBASE. Finally.6. 3. In parallel to searches of electronic databases. Studies included in the clinical evidence review were also screened for economic evidence. Studies that did not meet the inclusion criteria. Details of the search strategy for economic studies on interventions for people with depression are provided in Appendix 12. CINAHL and PsycINFO) were searched. OHE HEED was searched using a shorter. the general strategy for depression was used.000 references (stage 1). 1996) (see Appendix 13). 48 . This date restriction was imposed in order to obtain data relevant to current healthcare settings and costs. The systematic search of the literature identified approximately 35. reference lists of eligible studies and relevant reviews were searched by hand. The abstracts of all potentially relevant publications were then assessed against a set of selection criteria by the health economist (stage 3).2 Selection criteria The following inclusion criteria were applied to select studies identified by the economic searches for further analysis: ● Only papers published in English language were considered. were secondary publications to a previous study. with the final search performed in December 2008. all papers eligible for inclusion were assessed for internal validity and critically appraised (stage 6). or had been updated in more recent publications were subsequently excluded (stage 5). Full texts of the studies potentially meeting the selection criteria (including those for which eligibility was not clear from the abstract) were obtained (stage 4).

7 3. interviews from the Healthtalkonline website (www. 3. or if they utilised efficacy data that were based mainly on assumptions. cost–consequence analysis.healthtalkonline. Studies were excluded if they had a mirror-image or other retrospective design. Poster presentations and abstracts were excluded from the review. 49 . Methods and results of economic modelling are reported in the economic sections of the respective evidence chapters. a prospective cohort study.7.6. 3.4 Presentation of economic evidence The economic evidence identified by the health economist is summarised in the respective chapters of the guideline. Studies were included provided that sufficient details regarding methods and results were available to enable the methodological quality of the study to be assessed.3 Data extraction Data were extracted by the health economist using a standard economic data extraction form (see Appendix 14).6. following presentation of the clinical evidence. Full economic evaluations that compared two or more relevant options and considered both costs and consequences (that is.1 METHODS FOR REVIEWING EXPERIENCE OF CARE Introduction The chapter on experience of care (Chapter 4) presents three different types of evidence: personal accounts that were collected by the service user and carer members of the GDG. or a systematic review and meta-analysis of clinical studies. are provided in Appendix 15. Selection criteria based on types of clinical conditions and patients were identical to the clinical literature review. This criterion was in line with selection criteria from the previous guideline. 3. The references to included studies at stage 5 of the review.org). and provided that the study’s data and results were extractable. cost–utility analysis or cost–benefit analysis) were included in the review. and review of the qualitative literature. as well as the evidence tables with the characteristics and results of economic studies included in the review. this criterion was not applied to studies reporting utility weights that could be potentially used in cost-utility analysis. costeffectiveness analysis. However. Studies were included if they used clinical effectiveness data from an RCT.Methods used to develop this guideline ● ● ● ● ● Only economic evaluations conducted in the UK were selected so as to reflect healthcare resource use and unit costs directly relevant to the UK context.

Each individual signed a consent form giving permission for their account to be reproduced in this guideline. In approaching these individuals. and sections of the text were collected under different headings using a qualitative software program (NVivo). Two reviewers independently coded the data and all themes were discussed to generate a list of the main themes. and themes were identified.7.Methods used to develop this guideline 3. All individuals who were approached to write the accounts were asked to consider a number of questions (see Chapter 4) prepared by a service user and carer topic group4 which oversaw this aspect of the guideline work. Reviews were sought of qualitative studies that used relevant first-hand experiences of people with depression and their families or carers.3 Interviews from Healthtalkonline Using the interviews of people with depression available from healthtalkonline. and through various agencies with access to people with depression.2 Personal accounts The authors of the personal accounts were contacted primarily through the service user and carer representatives on the GDG. The personal accounts were read again by the service user and carer topic group. 2006). Any changes made for clarity were approved by the authors of the accounts. The GDG did not specify a particular outcome. Instead. 3. The full text of the accounts is reproduced in this guideline. if necessary. and the review team. 3. the review team analysed the available data and identified emergent themes. or where it was thought that further information would be helpful. The evidence is presented in the form of these themes. 4The 50 .7. the review was concerned with topic group comprised three service user and carer members of the GDG and two members of the NCCMH review team.4 Review of the qualitative literature A systematic search for published reviews of relevant qualitative studies of people with depression was undertaken using standard NCCMH procedures as described in the other evidence chapters. the authors of the accounts were contacted again if parts of their account were unclear or ambiguous. All personal accounts were read by the members of the service user and carer topic group. Each transcript was read and re-read. the GDG attempted to assemble a range of individual experience that reflected what the GDG considered to be important aspects of the care and treatment of people with depression.org.7. The methods used to synthesise the qualitative data are in line with good practice (Braun & Clarke. and the review team. These themes were developed and reviewed by the topic group and then incorporated in a combined summary with the evidence from the other two sources below. with selected quotations from the interviews.

and the guideline updated as appropriate. NICE then formally approved the guideline and issued its guidance to the NHS in England and Wales. Following the consultation period. 3. which were again developed and reviewed by the topic group.9 VALIDATION OF THE GUIDELINE Registered stakeholders had an opportunity to comment on the draft guideline. people with depression and companies have contributed to and commented on the guideline at key stages in its development. The evidence is presented in the form of themes.5 From evidence to recommendations The themes emerging from the personal accounts. the qualitative analysis of the Healthtalkonline transcripts and the literature review were reviewed by the topic group. The GRP also reviewed the guideline and checked that stakeholders’ comments had been addressed. all comments from stakeholders and others were responded to. the GDG finalised the recommendations and the NCCMH produced the final documents. 3. 3. 51 . These were then submitted to NICE. They are summarised in Chapter 4 and this summary provides the evidence for the recommendations that appear in that chapter. Stakeholders have been involved in the guideline’s development at the following points: ● commenting on the initial scope of the guideline and attending a briefing meeting held by NICE ● contributing possible clinical questions and lists of evidence to the GDG ● commenting on the draft of the guideline (see Appendices 4 and 5). Following the consultation. Stakeholders for this guideline include: ● people with depression/carer stakeholders: the national organisations for people with depression and carers that represent people whose care is described in this guideline ● professional stakeholders: the national organisations that represent healthcare professionals who are providing services to people with depression ● commercial stakeholders: the companies that manufacture medicines used in the treatment of depression ● Primary Care Trusts ● Department of Health and Welsh Assembly Government.8 STAKEHOLDER CONTRIBUTIONS Professionals.7. which was posted on the NICE website during the consultation period.Methods used to develop this guideline any narrative data that highlighted the experience of care.

● What possible treatments were discussed with you? ● Do you have any language support needs. 4. It should be noted that these accounts are not representative of the experiences of people with depression and therefore can only ever be illustrative. ● When did you seek help from the NHS and whom did you contact? (Please describe this first contact. There is then a summary of the themes emerging from the personal accounts. please explain how you gained access to services.1 EXPERIENCE OF CARE INTRODUCTION This chapter provides an overview of the experience of people with depression and their families/carers. please describe if you feel able to do so.2 4. having treatment and caring for someone with depression.Experience of care 4 4. The people who were approached to write the accounts were asked to consider a number of questions when composing their narratives. In the first two sections are first-hand personal accounts written by people with depression and carers. which appear in the final section.) What helped or did not help you gain access to services? If you did not personally seek help. did this have an impact on your receiving or understanding a diagnosis of depression or receiving treatment? ● What treatment(s) did you receive? Please describe both drug treatment and psychological therapy. which provide some experiences of having the diagnosis. including needing help with reading or speaking English? If so.1 PERSONAL ACCOUNTS – PEOPLE WITH DEPRESSION Introduction The writers of the personal accounts were contacted primarily through the service user and carer representatives on the GDG and through various agencies that had access to people with depression. the Healthtalkonline transcripts and the literature review.healthtalkonline. ● Was the treatment(s) helpful? (Please describe what worked for you and what didn’t work for you. which provides a basis for the recommendations.2. These included: ● When were you diagnosed with depression and how old were you? ● How did you feel about the diagnosis? How has your diagnosis affected you in terms of stigma and within your community? ● Do you think that any life experiences led to the onset of the condition? If so.org) and a review of the qualitative literature of the experience of people with depression. This is followed by a qualitative analysis of transcripts of people with depression from the Healthtalkonline website (www.) 52 . accessing services.

Experience of care ● How would you describe your relationship with your practitioner(s)? (GP/community psychiatric nurse/psychiatrist. employment and making relationships) and the lives of those close to you? Each author signed a consent form allowing the account to be reproduced in this guideline. low self-esteem and my underlying childhood issues. I was also on an antidepressant and tranquilliser for instant tranquillisation whenever I felt miserable.2. and the relationship between people with depression and professionals. please describe these strategies. so I had to learn in therapy how to deal with my own emotions from scratch. the negative impact on daily functioning. Initially I had 3 years of gestalt therapy with a wonderful therapist who came recommended by a friend. I then had psychodynamic psychotherapy for 4 years (while I 53 . each year becoming more confident. to talk to him for 10 to 15 minutes every week. concerns regarding stigma and discrimination in the workplace. all of the personal accounts received were from people who have/have had severe and chronic depression. ● Did you attend a support group and was this helpful? Did any people close to you help and support you? ● How has the nature of the condition changed over time? ● How do you feel now? ● If your condition has improved. The depression passed within 4 to 5 months. He enabled me to go to see him whenever I wanted. 4. spanning many years. for example private treatment? If so please describe what was helpful and not helpful.2 Personal account A I was 23 when I was first diagnosed with depression. Seven personal accounts from people with depression were received in total. my first experience of suffering with depression was most probably as a teenager. ● In what ways has depression affected your everyday life (such as schooling. living in a chaotic household with a parent with alcoholism and a narcissistic personality disorder. do you use any strategies to help you to stay well? If so. the need to take personal responsibility for and understand the illness to improve outcomes. mood swings and cruelty. paid and unpaid employment as an important part of the recovery process. 35 when diagnosed with major depressive disorder and 43 when diagnosed with dysthymia. However. and so on. I always think of the GP fondly as a life saver. Although the questions were aimed at people with any form of depression. During my childhood I had had to deal constantly with my mother’s tempers. issues around diversity. For the next few years I used therapy to deal with my depression.) ● Did you use any other approaches to help your depression in addition to those provided by NHS services. The themes that are most frequently expressed in the testimonies include trauma or conflict in childhood as a perceived cause of depression. the need for long-term psychotherapy for people with severe and chronic depression. The first treatment I had was when I was 23 with a wonderful GP who told me he had had depression and a breakdown at medical school.

nearly took from me completely.Experience of care also ran a self-help group for women). but it took a long time to heal. I found this psychotherapist from the UKCP list. But I kept quiet and got on with it. This changed over time. I also had a wonderful GP in 2002 to 2003. I was swimming and eventually was able to pay for my own psychotherapy. ashamed and humiliated. This left me feeling very distressed. and I had been open with my friends about being depressed and in hospital. Just getting off the huge amounts of medication was a feat I am proud of in itself. however. in addition to overcoming the depression caused by childhood issues and living a normal positive life which the medication. I had to give up work in 1998 to battle with it full time for a couple of years. but I don’t think I should have had to ‘re-prove’ myself as if I had been in prison. So I was just left with my depressed thoughts for 11 weeks. I have discovered through my own experience and working with service users. Overnight I lost two thirds of my friends and social contacts. I got a job in 2000. I came out and went back to work. Throughout this time I had battled with pervasive suicidal feelings and only my personal strength got me through. tranquillisers. to avoid the stigma. but my understanding of what hospital offered was not known to me. Also. and so on. to undermine and separate me from any compassion I could expect. family life was difficult and I had been working very long hours at work to solve all of these problems. although I was employed in the care environment. within my family. I was marginalised from external meetings for quite some time and my role was circumscribed. I learnt that it’s best to hide having depression. that it’s still best to hide having depression (or indeed any other mental illness) if you want to get a job and keep it. In my 30s. This has changed gradually over the years. I was given antidepressants. and gradually the major depression I had been in for 4 to 5 years lifted in 2002. I also didn’t realise that there was stigma around these matters. I had a major depressive episode and I booked myself into hospital which I now see as a big mistake as it was not therapeutic by any means. I have had two recurrences of major depressive disorder. At work. Subsequently. some people were not keen about me returning to work. 54 . not to mention the illness. I was offered no therapeutic help and I found the system of nursing within the ward very damaging – they just observed the patients and didn’t talk to us. It was sheer force of will that got me up and out each day. a mood stabiliser and benzodiazepines. I was tried on a series of drugs over a 7-year period: six different antidepressants and various mood stabilisers. Being on a level of medication that was unnecessary and toxic. During this period I also worked with teenagers and I found hard work to be a great help in having something to focus on and enhance my self-esteem. I had put on seven and a half stone since 2005 and I was threatened with high blood pressure and impaired glucose syndrome. but I could barely hold a conversation I was so drugged up. I had been having some housing problems. I knew that I was at danger point. my illness was exploited by my still-crazy mother. an antipsychotic. I begged to have psychotherapy but I now couldn’t afford to pay for it myself. who took it upon himself to (in his words) ‘have a go at’ at my consultant psychiatrist for half an hour on the phone about the cocktail of drugs I was taking. My GP helped me get off this cocktail of unnecessary medication.

but that with my hard work I have completely changed my life around and. It is worse if the negative attitudes are expressed by GPs and other medical practitioners. now agrees with me that psychotherapy. as I had previously done. This was even more wonderful in aiding my recovery and I had one session per week for a year working on my anxiety phobias. could be being turned away because of the lack of understanding. I would be dead now – and all because I have had depression. She is of the ‘old school’ and took a lot of convincing. she turned her ideas around about me and what I was able to achieve. narcissistic mother. nor the style of training they would want their therapist to have had. The psychologist was a wonderful professional who had faith in me and together we worked very hard overcoming the deep beliefs that I had held and which prevented me leading a full. My psychiatrist put this in writing to my GP in 2006. Where I currently live. making the most of my abilities and relationships in the present. you are treatment resistant. Choosing a therapist is as important as choosing a GP. I have always preferred a woman therapist. but at some point. especially those who are less articulate and bolshie than me. I am also learning self-analysis and skills building to enable me to keep an eye on stresses and challenges. and one psychodynamically or psychoanalytically trained. to self-manage and keep well. It is hard work but my personal stamina increases all the time. I can sometimes see a doctor’s face drop when they get to that point – some are not very good at hiding it. In 2006 I was turned away from a gastro clinic and told that my stomach pain and weight loss were because of depression and that the NHS couldn’t help me. 55 . The threatened ‘relapse’ has never happened. This enables me to build confidence and be the person I should be. My psychotherapist is helping me with positive attachment and parenting techniques to get to the point I should have been at. major depressive disorder is on my records everywhere. and forming a positive attachment in the psychotherapeutic environment. My psychiatrist. I have been having psychotherapy again since 2005. who I had from 1995 to 2005. in her terms. and it ‘woke’ me up. Even with lifelong illness you get one ‘go’ at it. Even now assumptions seem to be made when I have outpatient appointments for physical ailments because computerisation of records has meant even though I have recovered. well life. have been the best for me in my recovery. working on the final bits of damage done to me by my alcoholic.Experience of care Not being drugged up freed me and enabled me to function at work. patients cannot choose whether they would prefer a male or female therapist. If I had listened to that doctor in 2006. Within the NHS there is still a culture that if you don’t take any therapist. I feel quite sick thinking of how many people with depression and mental illness. I complained and the resulting CT scan showed I had cancer which when removed 6 weeks later was at stage 2. This therapy would not be available in the local mental health trust – there is only one course of psychotherapy available (1 year per patient). building my career and not being on any drugs. Stigma remains a problem however. and after an 18-month battle involving Mind and my psychiatrist. I got cognitive behavioural therapy (CBT) in 2004. not for any other reason. My self-esteem issues over my depression and weight had left me anxious though. She still confirms I was very ill. I am unlikely to relapse.

not mine. The counsellor was a very nice woman but I felt I was not being assessed. I am also conscious that any so-called emotional problems affect the way one is perceived and addressed. which I took for 4 months. I had already tried to help myself for 6 months and had bought many so-called self-help books. when I was 55. I understand the efficacy of exercise. I felt very much more unsettled at the end of each session than when I had arrived. I have a Master’s degree in social work and at one time taught counselling skills. in the summer of 2006. because of symptoms of what I felt was very severe and prolonged depression. made no difference to me at all. He offered no medication and no follow-up appointment. I was very reluctant indeed to seek help and many of my fears were in fact confirmed. Having said he could refer me to the mental health team. I got the impression she was talking about her own life. This is not entirely bad advice but much too crude. She talked a great deal about her upcoming wedding and for half a session explained the essentials of transactional analysis (which I’ve taught).2. who was better than the first but I could not afford to continue the sessions or to travel to see him. However. After three sessions I found another counsellor. positive thinking and relaxation. She explained that ‘if you haven’t had an adolescent rebellion you have one in middle age’ and told me to ‘get rid of’ people who were draining me. I was later even more annoyed when the difficulties with the counsellors were explained away by a mental health team worker as a disturbance of mine in facing the issues. I made an appointment with the private counsellor. I was able to change my GP. Again I found that the counsellor seemed to have a favourite model of human behaviour. He asked ‘are you depressed?’ I told him I was sufficiently distressed to consult a GP. Because of this. The major problem is that one cannot actually do these things when depressed and I believe those who have not been depressed cannot truly comprehend this at all. Fortunately. CBT. I sat in my car in the car park crying for 2 hours before I could drive home. Eventually I began a method of self-counselling: occasionally speaking aloud to myself in a deliberate effort to calm myself down since I knew that depression can be a result of over-stimulation. diet. I also felt that conclusions were drawn rapidly and inaccurately. I was working freelance as a trainer but no longer able to seek work and so I was without an income. transactional analysis.Experience of care 4. He told me to complete a ‘HADS’ test in the waiting room and put it under his door. although I was anxious about the cost. and so on. 56 . he said that they were ‘not very good’ and gave me a card for a private counsellor. She told me to keep a diary of angry feelings and never referred to it again. I am familiar with rational emotive therapy. I felt much worse afterwards knowing this and that I could not improve the situation. But I felt I had to try and help myself. The new GP provided much more help but unfortunately the initial medication (citalopram).3 Personal account B I first consulted my original GP in the spring of 2006. The GP whom I first saw spent more time looking at his computer than me. I had experienced a rapid series of distressing life events (a complex bereavement leading to feelings of alienation and isolation) and I had no support. person-centred therapy.

The escitalopram was beneficial and I have continued with it for over a year. a lot of people who are depressed have ‘abnormal appetite’. How can someone who is seriously depressed negotiate? I was also given the Aaron Beck tick box-type diagnostic tool which I found confusing. I left this appointment and began crying immediately – again I could not drive home for an hour. I still seem to need this medication. (For example ‘loss of appetite’ is difficult to answer. I also feel that I was quite poorly and was left to ‘wait’ to see if I would get better. but it was not even close to what I needed. The one friend who knows about my condition then took me to the surgery. which was frustratingly puzzling to me and based on no knowledge of me whatsoever. The registrar changed my medication to escitalopram. Even then I felt guilty for someone seeing me ‘as an emergency’ and I felt very bad about that. He said. I did very little talking and I could not summon the energy to constantly correct the assumptions being made which. I think that I expressed the issues quite clearly in the limited time. I still needed help. however. So then appointments had to be made a few days ahead. I was deeply grateful as my GP had kept telling me to continue the citalopram and wait for it to take effect. He was. I took extra medication to try and cope. which was one of the worst times for me. I found it disturbing to have to talk to a stranger yet again. But one can only go to the surgery when one feels capable of doing so.Experience of care My new GP referred me again for counselling at the surgery. The sessions often ended with an emotionally laden question or the advice given was more appropriate for a much older bereaved person. In the end it was a registrar GP who saw me in this deeply distressed state. ‘Yes. One needs to be able to access help when one needs it during the bad times. I now know that I was quite seriously ill at this point. However. In February 2007 I got into a very distressed state but could not get an appointment with any GP although I phoned the surgery four times.m. Appointments had to be made on the day at 8. I feel that getting the medication right and promptly at the virulent stage of the depression is vital. and my GP was offering what was available so I felt I had to accept it.) I find these tools very simplistic. My own reasonably sound knowledge of counselling actually seemed to be a disadvantage to me and I had to learn to keep quiet.30 a. Prior to my mental health team assessment interview in May 2007 (the GP registrar I saw in February had written again to the team to ask for an early appointment) I was in a very foggy state and was particularly vulnerable. I called the mental health 57 . again. seemed based on the counsellor’s own life. said he was trying to clarify why I was there and at one point told me I looked ‘alright’. I attended just a few sessions and then decided that this was a waste of resources. I felt that if someone would just skilfully listen and question (as I thought good counselling did) I could sort things out myself. The interviewer described himself as a nurse. had very little external support. which was very unhelpful. I’ve had bereavements too’ and ‘I don’t know why you have been referred’. quite good and he referred me again to the mental health team. I quickly lost confidence in my interviewer. There was a waiting list: I attended the first session and then there was a gap of some weeks (which was at the end of 2006). He also told me I had to ‘negotiate’ if the counselling is not right.

This meant that I had problems getting my pension (money problems started to become a major factor when my savings diminished). but has only so much time. She did provide details of a freelance psychologist. The occupational health professional said I had to have a consultant diagnosis. but it was almost a year before I could see a psychiatrist for a formal diagnosis. Over the past 2 years I have had to share my personal details over and over again with about 12 strangers. Up until 6 or 7 months ago I was feeling as if in a parallel universe. which was the one life-threatening condition which I had. The psychiatrist saw me for two sessions but explained that she could not see me again (as this was. My GP has done her best. there is no other help available. I had to miss one of the six sessions because I was not well enough to attend. While perhaps dismal. and was finding it very hard to concentrate on what was being said. The more contacts I had. Also. very expensive). I felt no ‘sense of achievement’: a lack of achievement is not one of my problems.Experience of care team and was told that I was bound to get upset ‘as I was talking about upsetting things’. With every other (physical) condition for which I have been referred I have been seen by a consultant at least once. I saw this psychologist twice paying £75 each session but just could not afford any further sessions. My GP had said that she would be able to refer me to a psychologist but that first I had to be referred to the mental health team. I was never clear about the role of the mental health team or what the ‘variety of options on offer’ actually was (in fact other than counselling there was ‘nothing else available’). I at last saw a consultant psychiatrist privately in January 2008. the problem is presented as being because of the vulnerability of the patient rather than the competence of the interviewer. I felt that my self-report was not being taken seriously and I was very confused about how I could present myself to make myself understood. I told my GP about this but she said I would feel a sense of achievement if I continued to drive! This greatly concerned me. I found this very disappointing and also embarrassing. 58 . I have had no further treatment other than the medication. I expect. it was a relief to have the diagnosis and it does validate my experience. But with a mental health problem. At one point I stopped driving as I knew that I was not safe to do so. akin to being in another universe. As my GP said very recently. It was not recognised that I was in a deep fog. I could not ‘wake up’ from dreams. and at one point as if I was living under water. half of them doctors ‘assessing’ me. and very unusually for me I could not get up until 10 am on some days. It was like offering aspirin for appendicitis. Again. I felt profound grief. but told me that I would have to see her privately. She diagnosed me with post-traumatic stress (I had been severely bullied at work before I left 10 years ago) leading to severe depression. which my former employer paid for. the more distressed I felt. When after many weeks I got to see the mental health team counsellor in June 2007 she told me the sessions were for 6 weeks so I knew immediately I could not be helped in this short time: I was taught ‘relaxation training’ which was inadequate for my needs. just ‘short fix’ stuff. I was going to have to tell yet another person about my life. I was referred by a GP and seen by a nurse (who thought I ‘looked ok’). and one wants to be a ‘good’ patient.

I spent 6 days there and was medicated. One problem was being not being able to work. The main complication was that my wife was expecting a baby and we were not getting on and constantly arguing.4 Personal account C Life experiences have definitely led to the onset of depression. I am mostly happier on my own though I am also gregarious and socially skilled. However. I had no friends and there was no support for my depression. Because of my past experience I couldn’t go to my parents or brother or sisters who lived near me. I do fear being destabilised by even small life events in the future as I know I am vulnerable and don’t manage such challenges well. I feel a little embarrassed that I do not have the things other people of my acquaintance have (family relationships and so on) and so I cannot talk the currency of that group (children and grandchildren). I was helped by seeing the consultant psychiatrist and I felt much better having been taken seriously. Because of my impaired sight I have had problems with sensitive hearing that made my life hell. The treatment was ultimately not helpful because there was no follow-up support. which was exacerbated by being sent away from home to be educated at a school for blind people. I felt like a prisoner and as if I was being tortured by everybody and everything with so much noise around me. My everyday life is affected as I am much less outgoing now.2. I also swim every day. Distraction helps if I can stop the circularity of thoughts. My own coping strategies are mainly avoiding known triggers. I was admitted to a psychiatric unit at the age of 30 because I was suicidal. After I lost my sight I felt I was rejected as a child and teenager by my family. or communicate with them effectively. I have been ‘let down’ so many times that I do not want to make the approach now. As the eldest of four children I bore the brunt of my father’s aggression and when I was older had to work in the family business for long hours and was punished at whim. relaxation 59 .Experience of care I now have far less faith in getting help so I do not know what I would do if things become worse. was having emotional breakdowns. I was sent to a local hospital by a doctor from the college and was diagnosed with problematic depression and was given more practical help than previously: I had some psychotherapy. and had too much to cope with at college. and there has been no support since then. This was due to a variety of reasons which had been building up to that time. self-monitoring and trying to get proper nutrition. They couldn’t give proper information in a manner that my family could accept or understand. I had an accident as a child which affected my eyesight and I have been visually impaired all my teenage and adult life. In 1992 I attended a college for the blind for training in the hope that I would be able to get a job. 4. But I am more accepting of my own isolation/difference from other people. I felt totally isolated and not wanted by anybody. Unfortunately this didn’t happen because I was so unprepared. Although I received a diagnosis of depression this was not fully explained to me and it didn’t do any good because ultimately the staff weren’t equipped to help me or my family. I felt totally lost.

I have been instrumental in setting up a local mental health drop-in centre and I am also a director of a local division of Mind and am standing as the BME representative on Mind Link. Nobody seemed to want to understand about my diagnosis and I didn’t feel I could talk to anybody because people are not equipped to provide support. My GP was a bit more helpful when I had my breakdown. society and family. for example. When I approached an Imam in a local mosque about a personal problem within the family I was told that religion would resolve it. There is a definite stigma towards mental health problems in my community. socially and financially. I had a severe breakdown last year and am concerned about relapse and was referred twice by my GP to the community mental health team. I am in a situation where I need the support of a therapeutic community or at the very least a safe place where I am able to get away from family pressures. It is more to do with faith and spirituality rather than religion. I feel like I am being pushed back. I have therefore spent the last 15 years working on complementary therapies and any improvement in my condition is due to the work that I have done. I have been rejected from services. So I think I am meant to live and survive – there is a purpose for me otherwise I would have given up long ago or gone to prison or got on drugs and alcohol. (I was able to access some CBT through Mind. I don’t have regular check-ups or practical support but I get help with medication and an occasional chat if I bring the subject up. My relationship with my current GP is better at the moment.) I have joined different groups. a bowls club for blind people.Experience of care classes and exercise for my neck. So I thank God I have not gone down those roads. I have a lot to offer despite no help being offered to me. I felt so bad I could have jumped off the roof. relaxation and emotional freedom therapy. I have felt like an outsider and have suffered rejection after rejection. He literally let me wander the streets. They believe in leaving it to the power of prayer. Many times I wanted to die and take the jump and I was saved. I was not seen by them. The self-help techniques I have used have included positive affirmation. But despite all this activity I am still disillusioned by the attitude of organisations that are meant to be dealing with mental health problems. I feel like my life is messed up physically. I feel like I am wasting my time trying. The only psychiatrist I have ever met told me that I would have to sort my problems out for myself. The CMHT did not do a good job of giving practical help: instead I was passed on to voluntary groups who were not fully equipped to offer support in a crisis or if I need help for referral from my GP to the CMHT again. which is Muslim. I have also received qualifications in holistic therapies. This was a very hard time and a struggle for me – both the college and the job centre rejected me by saying they couldn’t help me until I was stable. But this seems to mean nothing to them. mentally. He stirred up more trouble by visiting the family member with whom I was having difficulties. and in terms of work and education. It feels like a vicious circle: I have had a total of five breakdowns and have attempted suicide. But perhaps God saved me. I feel closer to God now and feel protected. and I have friends who have provided me with support. At the end of the college year I was advised to take a break of a few months. 60 .

He never gave any praise. He couldn’t work most of the time.Experience of care My feelings of alienation and isolation are exacerbated by family members who appear to have little appreciation of how difficult life is for me. I was constantly comparing myself to other people. My father also suffered from agoraphobia. I missed school in order to care for him or because he had hit me so hard I had a black eye and couldn’t go to school. 4. Depression has infected every part of my life. I felt as if I was always protecting my mother from my father. married young to get away from my father. It was a relief to know that my depression could be understood. which stemmed from my upbringing. I feel very isolated because my sensitive hearing makes me nervous and anxious in public places. I was first prescribed diazepam. I felt as though I was not as good as the next person. and to speak to someone who knew what I was talking about. I found it difficult to go outside and became agoraphobic. Because of my low self-esteem I couldn’t hold a job down because I felt as if I was not good enough to do anything. so I ran errands for him – I was his ‘skivvy’. I was on diazepam for about 6 to 9 months and then I came off it. and he never once said that he loved me or my mother. I felt at the time that life wasn’t worth living – I thought that practically it would be better to throw myself under a bus. which made me feel good because I was out of it.2. speed and barbiturates. I tried to look for a job but my feelings of inadequacy and paranoia returned: I felt as if people were looking at me and talking about me. It has slowed me down. When I started puberty I felt different from other people. if not treated. and I didn’t want to take anyone home as I was ashamed of my father. I knew my feelings were different from those of other people so I went to see the doctor by myself when I was 16. When I was 15 or 16 years old my father tried to kill my mother when he found out she was having a relationship with another man. 61 . I found it hard to learn at school and later I found out that I had dyslexia. The doctor knew immediately that I was suffering from depression. I couldn’t work but at least it was a lift and that is what I felt I needed. led to loss of self-esteem and made it difficult for me to get work. my siblings and me. If I hadn’t gone to the doctor I would have killed myself. I came from a poor background – my father was diagnosed with bipolar disorder when he was in the army during the Second World War and after being discharged he spent a year in a psychiatric hospital. which affected my mother. Eventually I found a job I liked and when I was 18 years old I started having serious relationships. who are older than me. Both my siblings.5 Personal account D The depression started when I was young (I am now 57). I was still living at home then and stayed to protect my mother as my father was still beating her. There were a lot of kids at school living in poverty but life with my father made me feel very inadequate. I was prescribed one tablet a day but I took three or four. My father had bad mood swings. Nothing else was offered to treat me so I treated myself by using cannabis.

She was hospitalised and I stopped working and looked after the baby – it was like being a oneparent family. I was trying to keep it together but she believed I was having a nervous breakdown. I went to see my GP a few times during this time and they were sympathetic to what I was going through. I battered myself with questions: what is the matter with me? I was consumed with all the thoughts of what had happened in the past. everything had been completely obliterated by illness. it seemed like there was a massive void. this was just a mask. My wife’s illness made me feel depressed but I couldn’t show it. I felt like I had never had a Dad and I became very good friends with a man in his 60s who I tried to adopt as a father. I needed some emotional help and I needed someone to talk to. After my wife had sufficiently recovered from her first episode of schizophrenia (it took about 9 or 10 months). I also began to have panic attacks. The counsellor was better than the antidepressants. For a few years I was in a cycle of relapsing and recovering – I was up and down like a yo-yo. By the time I was in my early 30s I was working in the building trade as a site manager and I was earning good money for the first time. I was working flat out and didn’t have time to think about myself. I felt as though I had lost my wife and there was just a shell of a person there who used to be my wife. I loved my wife and that was enough in life. Although I was convincing my wife that I was coping.Experience of care I finally left home at age 21 when I got married. I was happily married and away from my father and it felt like depression was behind me. I stood by his grave and I couldn’t cry. I felt numb about it all. There was no time for myself and I stopped communicating with people. Children completed the marriage. I was offered antidepressants but I didn’t take them as I didn’t want my wife to see them. This lasted for a few months before I began to feel low again. she was hearing voices and it was as if the gates of hell were opened and everything came out. In the following year my wife was diagnosed with schizophrenia. I felt as if life was taking off. The illness was like a bereavement. But my wife was feeling better and we wanted more children so the doctors took her off her depot antipsychotics and antidepressants. I felt as 62 . The asthma hit me hard as I was my wife’s carer and I looked after the children. I thought about what it had taken out of me and I would sink into depression and phone up the Samaritans. My personality only grew when I got married. Throughout her illness I was on an adrenaline rush. When she became pregnant she was happy and like she used to be before the illness. Shortly after this I was diagnosed with asthma. I couldn’t set a course for a life. In the end I took time off work. It gave me a good lift. I was a machine trying to keep my family together: looking after my wife and kids and working. I realised how badly it had affected me. I felt that there was a crater in my life where my father should have been. which was considered by my doctors to be my major illness rather than depression. She was 28 at the time. My Dad died in 1983. I was determined not to be like my father and I appreciated what I had. I started taking amitriptyline and I also saw a counsellor for 3 months. I didn’t have anyone to look up to – no one to build a personality around. In 1987 my youngest son was born but 4 months after his birth my wife became very ill.

that my purpose was to make someone else become well.2. I got depressed about what was happening to my son because I didn’t want him to go through the same things that his mother and I had been through. To try and cope with the symptoms I grin and bear it or I try doing something different – getting away from mundane routine. To be honest. I don’t want to die in between the bleach and the biscuits. Although people think that I am stable. for me. I recognise that I will never be free of depression but as I get older I understand more about it.6 Personal account E I was 27 years old when I was first diagnosed with depression. but CBT is not the right treatment for everyone with depression and this needs to be recognised. If I feel like this for more than one day then I start to worry and I know I am depressed. it’s like an old nemesis. I am particularly interested in the political side of how people with mental health problems are treated. 63 . He was badly bullied at school for having a mother who was a ‘nutter’ and got very depressed. I hate taking tablets. I become more depressed when there is a crisis – and there always seems to be a crisis in my family. In 1997 my wife relapsed again and it affected our youngest son very badly as he had not seen his mother this way before. But I have accepted my depression as I have lived with it for so long. I don’t want to kill myself.’ This happened several times around this period. I just didn’t know it at the time. 14 years ago. I believe that my depression was caused by my childhood experiences. Then one day I was pushing a trolley around the supermarket and I thought ‘I don’t want to die in a supermarket. When he was 15 (in 2003) our son was also diagnosed with schizophrenia. Eighteen months ago I was taking venlafaxine but I am not currently been treated for depression. I can now recognise when I am becoming depressed. is like a form of counselling. which. I got involved with voluntary groups when my wife got schizophrenia: I am the chair of one voluntary organisation and I work for another. I didn’t go to doctors as I thought they would think I was nuts. It’s a waiting game. but depression is such an individual illness – it has got many different faces and it can be caused by many different things. I care for both my son and my wife and I will never turn away from them. When I was first ill I thought I was a lunatic because I was taking tablets. 4. and I do a lot of media work. If I do need help I find that counselling is best for me. Therefore should people with depression be treated in the same way? I am encouraged to see that a lot of resources are being put into providing CBT for people with depression. I did not see that there was something wrong with me. It’s a part of me.Experience of care if I had become invisible. although I have not seen a therapist for a few years. I think I started to get depressed 6 years prior to diagnosis. I get black days when I wake up in the morning and I am totally unmotivated and I couldn’t even care if I won the lottery – it would make no difference because I feel so lousy. The horrid feeling of not being as good as other people is not there now because I feel that I am helping. I am now able to talk to my wife about being depressed rather than trying to hide it from her and I talk to lots of other depressed people.

I was offered counselling and/or medication. so was given counselling immediately. and was physically and emotionally cruel to them. How could I tell people that I had spent the week trying not to kill myself. I was also prescribed an antipsychotic drug. in terms of therapeutic input it did nothing – people would talk about their week and how awful life was. though time still went slowly. I attended and it helped much more than I realised at the time in that I formed friendships that were very supportive. Without doubt. but I felt like a zombie and could not look after my daughter. I established a good and trusting relationship with the counsellor who helped me to understand what was happening to me. My Dad remarried a woman with three children. I had gone to the doctors knowing something was wrong. but I couldn’t accept at that time that depression would be part of my life forever: I found it difficult to listen to others about how they were managing their lives living with depression. ending a 6-year relationship with my boyfriend. but not knowing what it was. but I could see no other way of stopping the pain. Of the professionals listed above. My mother died when I was 5 and after that my two younger brothers and I were not allowed to talk about her. I did not feel any better. My Dad hated her children. so did not take it often. However. when that was all I wanted to do? It was not that I wanted to die. and was physically and emotionally cruel to us. though I didn’t tell family. being made redundant from my part-time job. and then being physically assaulted. who I saw for around 18 months. but I was in no fit state to engage in any meaningful therapy. I plummeted further. as it would make me feel better more quickly. but I couldn’t do that. Reading books about depression and self-help gave me an understanding of what was happening to me. I was seeing the psychologist. without doubt the CPN helped the most. One of my stepsisters sexually abused my youngest brother and me. and my stepmother hated my brothers and me. there was none. my childhood experiences have also contributed to a life of depression. run by my counsellor through the NHS. As well as the treatments listed above. but it was not long before my Dad and stepmother hated each other. because they wouldn’t have understood.Experience of care At first. I had a good relationship with her. I was relieved at the diagnosis. I saw a psychologist for the following 18 months. A month or so after starting medication. When I was at my most depressed. I knew that this ‘breakdown’ occurred due to the events that had happened the previous 18 months: the sudden deaths of two close friends and my grandmother. On one occasion I went to a voluntary agency support group. and was seen by a psychiatrist who allocated me a CPN. and it was unbearable. Depression filled every second of every minute of every day. However. I was fortunate in that I was able to sleep a lot (up to 15 hours a day). until I was able to slowly start rebuilding my life. I thought I was going to get better and it would never come back again – how naïve was I? 64 . When my ‘time’ was up seeing the counsellor. I knew that I had to have medication. as I was too ill. and were physically and emotionally cruel to each other. I had already withdrawn from my friends and community (due to the depression) so in terms of stigma. while I was having counselling I was told that I should attend a women’s group.

I went through my work with him. For years when I was depressed I needed to sleep a lot and I also put on weight. He advised me that I could take time off of work. as it did not get to the root of the depression. he would get someone else to do. I didn’t recognise I was depressed for a long while and by the time I went to see my doctor. I feel frustrated that there are no services available to me now. Now I struggle to sleep (which has its obvious disadvantages) and I tend to lose weight. I do not see him often.Experience of care Over the years. and was able to do everything because he took the pressure off me. by the NHS. due to the psychodynamic therapy I have had. could be taken away from me. I have been prescribed most of the SSRIs. I often feel that my life is hanging by a thread – that at any moment. he would go through everything I needed to do. He told me that if I felt unable to do something. Over the last 2 years I have paid privately to see a psychotherapist and had psychodynamic therapy. but some times are better than others. However. This suits me as it means I am under no obligations or pressure from him. though. I have had to build my life around periods of depression. Every morning for about a month after that. and I never took any sick leave. I did receive further counselling on one occasion. in addition to my full time job to pay for this. He told me to see him at any time I felt unable to do something. but because of his commitments. and so took 2 years to recover from. I function very well. he would come into my office in the morning to see how I was. but on a positive note it can ease quickly as well. my life. They worked to varying degrees. On one occasion I told him that I was going to have to take sick leave as I was very depressed and could not work. but it was not particularly helpful. there is always the fear that I will get too ill to work. But when it is severe. for which I am resentful. I have learnt to live with this. especially at work. Whereas now it can very quickly become severe. so long as it doesn’t interfere with one’s work. it was too late to treat successfully. but with the help of my GP (who has been very supportive and allows me to manage my depression my way) I have not had to say it is because of depression. I am currently seeing someone. but am aware of the limitations this poses for me. On the surface. it would be helpful to be able to access services 65 . I have an excellent manager at work with whom I can be honest. I have managed to qualify at university in the career I have always wanted. that I have worked so hard to build up. and I have had to get another job. and know that I am pretty good at it. rather like chronic back ache it is always there. Financially. There is a general acceptance at my place of employment about having depression. I have had to have the odd day/week off over the last few years. Depression for me has changed over time. This has been the most helpful in terms of trying to repair and understand the damage I experienced as a child. Depression is with me all the time. However. but that if I wanted. but the most distressing aspect for me is that they all seem to affect my memory and articulation. It is on this basis that I choose not to engage in a long-term relationship. I believe. no one would ever believe that I have depression as I am a good actress. and I love my job. this has been difficult.

which made me worse. I didn’t want to make relationships because I lost trust in people.2. on the NHS. Antidepressants and counselling were discussed as possible treatment options and I was referred for counselling but had to wait 18 months. Only my close friends knew that I had depression – I didn’t want people to know because there is very little understanding within my community. My family suffered as I was not really there for them and I couldn’t work because my illness was too severe for me to function normally. When I finally saw a counsellor. which was useless. He was brain damaged and I looked after him for 25 years until his death. 4. In the end I was put on Prozac which did help to improve my symptoms. I am now very seldom depressed. to a man who became a violent alcoholic. I’m just another number in a long queue. I married for a second time. However. I was offered hypnotherapy. which I didn’t want. I wanted counselling. My father then attempted suicide and was on a life support machine for 2 weeks. which can get to the root of the issues that cause depression.7 Personal account F I was first diagnosed with depression in 1999 when I was 44 years old and was feeling suicidal. We had four children but we could not provide them with much at Christmas and for birthdays. My relationship with my psychiatrist is non-existent. Depression devastated my life. At the current time I am still on antidepressants but I am ready to come off them. things have improved over the years. I have attended a Christian counselling organisation in the city where I live which has been brilliant. My doctor doesn’t have a clue who I am. I shut out a lot of people because I could not socialise when I was so ill. Because of the way I had been feeling I was relieved to have a diagnosis. I tried various medications. Because of his drinking he lost a lot of jobs because he was too hung over to turn up and we were often in debt and lived in poverty.Experience of care immediately from a team that knows me and can support me without me having to go through a series of assessments and then being told ‘well you can go on the waiting list for this service. I now know that I will have depression until I can resolve my childhood issues. I also feel that long-term psychodynamic therapy should be available. My mother died when I was 15 years old. We struggled financially to provide food and the basics. When I became suicidal I went to see my GP. My daughter was 3 months old at the time and I never got her back. such as Prothiaden. 66 . If I have any low moods I go back to my counsellor and exercise regularly and eat healthier food to stay well. There were well-trained counsellors available who were very supportive. but you can only have this service for a particular length of time’. He was very attentive and took me very seriously and referred me to a psychiatrist and a mental health clinic. After 9 years of being off work because of illness I am now getting back to work on a job placement. I was married at 18 and my first child was kidnapped by her father after I left him. The house became a tip. Two of the counsellors maintained contact in between appointments.

As my condition improved I started to research my illness online and also made online contact with others from across the world suffering from mental illness. and because of how I was feeling. Since being ill I have changed my GP four times due to moving around the area (one GP retired). I have been on medication and although no longer on lithium I feel that it is only over the last year or so that I have been listened to by my GP and psychiatrist. Their approach has differed. I have obtained such information from what I have discovered on the internet and from fellow service users and the voluntary sector. this antidepressant did not work and a few months later I returned to see my GP and asked to see someone. I was given no information about it from my GP. I was unemployed having been made redundant several months previously and also my marriage was in difficulties. I think that these things contributed to triggering my depression but neither was responsible in its own right. I feel this has slowed my recovery and has left me to deal with several issues that I feel could have been dealt with by a psychologist or psychiatrist.2. I could usually find someone somewhere in the world 67 . During the 8 years I have been ill. I isolated myself from my family. Initially I had three sessions with a psychologist who said that she could not help and referred me to a psychiatrist. I have found the internet very useful for getting information about my condition and when I was very ill and needed to talk. psychologist or psychiatrist. At first my GP was reluctant to do anything but after several visits she relented and prescribed me an antidepressant. Up until that point I had no experience of mental illness or knew anyone who suffered from it. The one real issue I have about my treatment is that over the 8 years I have only had three sessions with a psychologist and the rest of the time it has been purely medication. Other than that the only other contact I had was with the nurse who took blood samples to check my lithium levels. On reflection there were signs of problems a couple of years previously. such as side effects. Also it concerns me that I was never offered any help or advice on managing my condition. He changed my antidepressant and I then saw him on a monthly basis. and has often been inconsistent. Fortunately my wife at the time had accompanied and backed me up otherwise I don’t think the GP would have referred me to a psychologist/psychiatrist. my marriage broke up.8 Personal account G I was first diagnosed with depression in 2000 at the age of 42. and it is only my most recent GP who I feel has listened to me and worked with me dealing with any medical issues around my condition. Once my condition had stabilised the only contact I had with my GP and psychiatrist was to either get my prescription renewed. Unfortunately. or seeing my psychiatrist every 3 months for 10 minutes.Experience of care 4. Eventually I was prescribed a mix of a tricyclic antidepressant and lithium carbonate that proved more effective at controlling the symptoms. The diagnosis was not a surprise as it had taken a few months for me to decide to go to see my GP as I tried to cope with it as best as I could. during which time I was unable to work. However this took 18 months. At the time I was diagnosed. This second antidepressant did not work and it was changed again. I think that was the reason I isolated myself from my family more and more as time went on.

challenging to one’s own physical and mental health. the questions included: ● How long have you been a carer of someone with depression? ● How involved are/were you in the treatment plans of the person with depression? ● Were you offered support by the person’s practitioners? ● Do you yourself have any mental health problems? If so. It can be frustrating. I must say that caring for someone is one of the most rewarding things I have done. but ultimately helping someone make the most of their lives by helping them in their most vulnerable moments.1. I didn’t have to. I’ve also done voluntary work for 18 months with a variety of organisations involved with disability and mental health. is rewarding. In October 2002 I went to university as part of my ‘recovery’ graduating with an MSc in 2003. 4. I still have some issues due to the depression. but for carers of people with depression. and so on) ● Did you attend a support group and was this helpful? Did any people close to you help and support you in your role as a carer? ● In what ways has being a carer affected your everyday life (such as schooling. Also keeping my mind active helps and doing voluntary work gives me a feeling of having ‘value’ in society. it is noticeable to me that my mood is more variable than when I was on lithium.2. but know that it will take time to resolve these so I try not to let this affect me. were you offered an assessment and treatment by a healthcare professional? ● How would you describe your relationship with the person’s practitioner(s)? (GP/community psychiatric nurse/psychiatrist. Although I feel well at present. 4.Experience of care to talk to 24 hours a day. since then I have found it difficult. The other advantage was that when I didn’t feel like talking. The biggest effect depression has had on my life is when it comes to employment.2 Personal account H Firstly. Over the years I have formed an online network of fellow sufferers and we keep each other up to date on anything of interest happening in the various countries regarding mental illness and its treatment.3.1 PERSONAL ACCOUNTS – CARERS Introduction The methods used for obtaining the carers’ accounts was the same as outlined in Section 4. employment and making relationships) and the lives of those close to you? Two personal accounts from carers of people with depression were received. 68 .3. but the strategies I have in place help me cope with this. exhausting.3 4. Although this did not help me find work I found it very beneficial to me in that it kept my mind active and this is something I have continued to try and do since then. Since being diagnosed I have only worked for 8 months in paid employment. Although I did not have a problem getting work before being diagnosed.

at present he has been ill since 2005. organise things and occasionally exert pressure to get him out of bed and even out of the house. as I can see he desperately needs to be helped with changing his thought patterns to positive thoughts. It actually has the good effect of getting him out of the house at least once a day. so it’s an uphill struggle. As a result of this illness. and though I wouldn’t say this to him. the pressure of his overwhelmingly negative thoughts and depressed ways of thinking can be a burden. I have neuropathy and sometimes this overwhelms me and I have to lie down for a couple of days to let it ‘wear off’. I also emotionally support him by listening. and that these things should be included in his care plan. but the pressure of 24-hour depression wasn’t doing me any good and I had to move house to be able to care for him again. I see him two or three times a day at either his home or my home. He doesn’t want to think about bills and money. But most of all I provided practical and emotional support. If it goes on too long. My partner is able to get my shopping and visit me and strangely this seems to take his mind off his own suffering for an hour or two. As his carer. Most carers I have met also say this. as a way of trying to create positive thoughts in his repertoire. and runs up huge phone bills when he is depressed. we can’t live together anymore. because childhood is when we should be able to expect to be nurtured ourselves. But I think it can be damaging for children to care for an adult without support. I have struggled for 2 years to try and get him CBT without success. and sometimes I get angry with his social worker. It’s a constant battle to not get services withdrawn. I attend his reviews and make sure he is looking after himself as regards to diet and exercise. I am trying to get him a review of his medication plus a referral to an occupational therapist for support around physical exercise. he gets cross. It’s hard for me seeing him suffer. 69 . it has changed the dynamic between us forever. because sometimes he would rather sleep 18 hours a day every day. as a carer. I plan trips out. I then became a carer to my partner. trying to get them to take this seriously because his father had two strokes at his age and he himself has been warned about fat around his heart. I am more like a mother than a partner. My partner has had two long periods of depression. His best male friend and I have decided to only respond to positive subjects that he brings up. when they can’t see that physical health and other risks are associated with his depression. and trying to encourage him to be positive. I battle with his doctor and social worker over this. In a way. to come and see me. and wants me there to support him. though. I have to constantly nag him to get him to try and keep an eye on his expenditure as it is a risk to his welfare. I was my mother’s carer when I was a child and teenager and I made sure she ate properly and took her tablets.Experience of care This applies to any caring. At one point last year he hadn’t seen a social worker or a housing support worker for 3 months. which would help his overwhelming depression. as he still has physical strength. They have tried the newer antidepressants on him but one of the old favourites seems to be doing the trick. His physical health is suffering as a result of extreme weight gain because of the medication and a lowering of his activity levels both because of medication and depression. working through problems with him.

or telling her about school with funny bits to try and make her laugh. Or telling my Nan and Grandad about how she was so they could come and help.3. mostly regarding carer. caring is rewarding but it can also be tiring and frustrating. but now it’s more. But I wouldn’t be without her or want to leave her on her own – she’s my Mum! I try and be positive and jokey. stopped their antidepressants at one point and crashed into a psychotic depression. very important to most carers: I have heard other carers say that they go to work to get a rest from the overwhelming nature of caring. and his employers have given him until December 2009 to get through this depression. make sure I get in straight away from school because I worry about her when I am out. like – we do normal things then and she’s the normal bossy Mum. I am more of a grown-up than when she’s well. nag her to get dressed when she’s depressed. 70 . I do well in school. When her friends come round and take her mind off it for a while.3 Personal account I My Mum has been depressed on and off since I was a 7-year-old boy (I am now 15) and I have been caring for her since then. Work is very. I hope she gets better soon. 4. and. I get her tablets. The role of being a carer for someone with severe depression has added to my own symptoms of dysthymia over the years because of the sheer pressure of coping with someone who turned down treatment. I wish she was like other Mums sometimes. or who have a previous history of depression. I know the signs. well. but I am able to keep myself busy and to have time for myself through work. I go out and do usual things too so that she doesn’t worry about me. and normal. make appointments. and make sure she sees her therapist even when she doesn’t want to go out and sometimes get her friends around for a surprise to make time pass for her.Experience of care When my partner was depressed previously. all the time. I sit down and talk with her. This means I have the time to care. I was able to support him and get him back to full time work within a year. She’s not depressed all of the time. Carers who become ill with depression or anxiety. and answer the phone. and it nearly broke me. and working from home in the mental health and housing fields. she laughs. should be offered support. I have built my career around being self-employed. As I have said. but I know it is a real risk for him and not working in the long run would not help his self-esteem. sort out food shopping. Now he has been off work since 2006. My Mum takes tablets and sees her therapist but I think seeing people really helps her. Then she goes quiet and stops going out and seeing her friends and I try and cheer her up and make things better for her. This was a huge burden and local services left me to cope with this on my own 24 hours a day. I go to my room when I feel cross and sometimes talk to my friends. When I was small it was just making her a cuppa now and again. Mostly she’s well but now and again she gets depression. and it’s fun when she’s well. behave myself and be there for her. resident or service user issues at strategic level.

and sections of the text were collected under different headings using a qualitative software program (NVivo). The anticipated headings included: ‘the experience of depression. and I meet other people like me caring for their parents. I play pool and we have days out – we went to Alton Towers which was fun. have pizza.healthtalkonline.org).4. fool about. but we don’t talk about it all the time. The same transcripts were also reviewed by Ridge and Ziebland (2006). The methods adopted by Healthtalkonline to collect interviews were two fold.4 4.4. Healthtalkonline provides interviews with people with both physical illnesses and mental health problems. Sometimes we watch films. We want to get away from it just for a few hours. narrative dataset. 71 . Each transcript was read and re-read. she can’t help being depressed. People sometimes think or say my life is sad. but I know it’s not my Mum’s fault. and there’s a support worker if you do want to chat. which is included in the review of the qualitative literature below. I love her and where else would I want to be? She helps me too. From the interviews. Second. It’s good meeting other young people like myself who are carers too.Experience of care Don’t forget your friends when they are depressed. And chocolate sometimes helps too! For a while I had no support but now I go to the Young Carers’ Centre in our town. the participants were typically asked to describe everything that had happened to them since they first suspected a problem. The review team undertook their own content analysis of the interviews to explore themes that could be used to inform recommendations for the provision of care for people with depression. Two reviewers independently coded the data and all themes were discussed to generate a list of the main themes. the review team for this guideline identified emergent themes relevant to the experience of people with depression that could inform the guideline.2 Methods Using the interviews available from Healthtalkonline. The review team decided to undertake their own analysis to cover a wider range of themes than those focused upon by Ridge and Ziebland. First. to obtain a relatively unstructured. I say.1 QUALITATIVE ANALYSIS Introduction The following section consists of a qualitative analysis of personal accounts of people with depression using Healthtalkonline (www. be normal. I had a carer’s assessment there too. The researchers tried not to interrupt the interviewees. 4. the review team analysed the experience of 38 patients from across the UK. a semi-structured interview was conducted in which the researcher asked about particular issues that were not mentioned in the unstructured narrative but were of interest to the research team. 4.

information on issues that are particularly pertinent for people with depression that could be used to inform recommendations may not have been collected. being neglected and isolated and being treated different academically. Many people described the death of a family member or friend as a trigger of their depression. the momentum of the skin of the balloon would just push me back in. I think. very thick balloon and no matter how hard I pushed out. It [depression] was like being inside a very. For example. some people described life events which they felt had caused the disorder. isolation and feeling withdrawn. one person described having a ‘racing’ mind that was ‘zooming into miserable places’. 4. probably about 4 or 5 months after starting my first year. . ‘accessing help and getting a diagnosis of depression’. they also described feelings of loneliness. I couldn’t feel anything for the children. being bullied . . Some people commented that stressful situations at work contributed to the onset of their depression. I think they all combined with my lack of social skills. body aches. . . using Healthtalkonline did highlight issues regarding depression that can be reflected upon for the purpose of this guideline. There are some limitations to the qualitative analysis of people’s experience of depression and its management undertaken for this guideline. One service user summed up various life events that she believed were associated with her current state of depression: All these experiences from earlier on in life. . which I’d not had a chance to develop until that point when I got to university .4. Some people used metaphor and allusion to illuminate their experience of having depression. Others used analogies such as depression being like a ‘brick wall’ or ‘being inside a balloon’ to describe how depression can act as a barrier from experiencing the world: I couldn’t feel anything. ‘stigma and telling people about depression’ and ‘electroconvulsive therapy’. I was just feeling very low and very lonely. .3 Experience of depression In recounting their experience of depression. ‘pharmacological interventions’ and ‘healthcare professionals’. . within a few months . tearfulness and sleep problems. Moreover. I did become very depressed.Experience of care ‘psychosocial interventions’. my Mum dying. Other people listed the symptoms they were experiencing: lack of pleasurable experiences. the review team did not have access to the full interview transcripts and therefore had a selective snapshot of people’s experience. Some of these events were childhood experiences including both problems in the family and at school. However. The headings that emerged from the data were: ‘coping mechanisms’. As the review team relied on transcripts collected by other researchers with their own aims and purposes. . 72 . needy . I couldn’t feel anything for [husband’s name].

These thoughts were described by people with depression as irrational and often caused them to jump to conclusions. having become more confident. you’ll always pull out the negative over the positive if you ever see a positive. oh. One person explains how she experienced negative thoughts: I call. For example. . if for one positive you’ll give ten negatives. Another common theme was that people felt that they appreciated life in a different way after having been depressed. And the doctor said to one of the nurses. ‘Go and get so and so . You know. Many people also felt that experiencing depression had made them re-evaluate their lifestyle and that this had led them to make some important positive life changes. understanding of others. I don’t want this. . ‘I wish you’d leave me alone. and I just thought. Some of the suicidal thoughts relating to suicide were: the ‘world would be a better place without me’. we’ve got about 10 minutes or he’ll be gone’. And with depression you see it really negatively. one person said: I can listen to music and appreciate it in a different way . and with a doctor and nurses around the bed. positive. you’ll . able to support others and able to do ‘something positive and . . People also described feelings of suicidal ideation and some disclosed their experiences of attempting suicide. ‘life wasn’t worth going on’.Experience of care A prevalent theme in the interviews was the presence of negative thoughts. or the only reason. One person described a suicide attempt: I can remember being almost unconscious. You see everything negatively. And I could hear him. and I have. . .’ However many people also identified positive aspects of having experienced depression. and whilst the depression wasn’t the main. that I left. . creative’. thank God I left there when I was 36 rather than 56. I understand that I need sort of time for me now. I’m warm and comfortable. Another person described the positive effects of having had depression: I think it’s [depression has] sort of made me question what I thought was good about my life because I was in a very busy and hard-working career. seeing things a particular way. . and ‘life was completely out of my control’. Basically it is my mind. it can move me now. Something on the TV can move me now. . I do think. and 73 . They also said that they had become more aware of themselves and their feelings and more able to cope with stressful events. for example. there was a reorganisation at my work. what I’ve got in my head my chatter box. I feel things and things affect me. One person described having had a breakdown as a ‘breakthrough’.

. . I’m OK I’m taking Prozac and then of course I knew quite a few people who were taking it as well. the right to have some quality time for me. they described their experience of receiving a diagnosis of depression. ‘Try getting more sleep. There was also stigma around receiving treatment for depression for both psychological and pharmacological interventions: It took a hell of a lot for me to go to therapy. . I’ve tried .’ [laughing] I was like. Once people with depression accessed help. B: therapy makes you a nutter.4. . my feeling is that really he should have asked a few questions and could possibly have diagnosed that I was depressed. And so I made an appointment with the doctor the next day. The majority felt that stigma still existed while a minority thought it was less prevalent than it used to be. it didn’t work . . yes. And that was a point where it was clear that . 74 . . And it doesn’t help. I couldn’t follow the sentences to actually read it out loud. . . so it was like ok like join the club. You know. . so hostile kind of lower middle class sort of feeling about that sort of thing. I’ve tried that. it’s like a happy pill isn’t it. . Some described how there is not enough recognition of depression and how often when they presented with sleep problems or loss of interest in sexual activities to their GP. You know A: nutters go to therapy. I could have thought of that. such as sleep deprivation and lack of concentration: I was putting my eldest daughter to bed and trying to read her a child’s story. And it was OK to say you were on Prozac. everything. these symptoms were not initially recognised as symptoms of depression: I went to the doctor and I said . . .Experience of care that I’m a person in my own right. 4. I no longer had the concentration to read . These were the kind of things that I grew up with.’ And he just said. . 4. and I actually found . I had to seek help. Conversely one person said it was quite ‘fashionable’ to be taking medication: Prozac is quite a fashionable antidepressant.5 Stigma and telling people about depression Some people described the stigma of having a diagnosis of depression. ‘I sleep but I always feel tired . . you know. and I’m important and I have.4 Accessing help and getting a diagnosis of depression Some people detailed how a particular event or problem prompted them to access help.4.

I think it’s the stigma thing . The older I’ve got. .4. . . . but after a while. and doing things . were needed to resolve the maladaptive behaviour and distorted thoughts that contributed to their depression: These tablets helped me . some people encouraged others to speak openly about their condition: You should tell someone now. the more I’ve found that it’s acceptable to say to people. I said. it can be a friend you know. I was in a psychiatric hospital for a month and then outpatients for a further month and now I’m at work part-time to try and get back into the swing of things slowly. I felt better [with medication]. instead of chatting to the brick wall. so it’s easier for me not to admit to that weakness. ‘I’m depressed at the moment’. And instead of it going round in your head and trying to sort it out. but I still didn’t have ways of dealing with things. Some described their experiences of telling friends and neighbours and stating that it helped them.Experience of care Due to the stigma surrounding depression. one person made a joke to ease the situation: I was just really outright. rather than medication. ‘I’m not loopy’ and he just started laughing. 4. . some people found it difficult to talk to other people about their condition: I can’t talk to my family about it.’ And he just looked at me . . my perception is that I would be seen as weak and not coping. I realised it sorted out my brain chemistry. because I’d just turned it into a joke. it doesn’t have to be the doctor or a therapist. ‘It’s ok though. ‘Ok. However. 75 . and I just said. . People with depression expressed the need for psychosocial interventions when the cause of depression was deemed to be psychological rather than a ‘chemical imbalance’.6 Psychosocial interventions People with depression discussed their positive attitudes towards psychological treatments: Sometimes you do need to talk to somebody who you don’t know.’ I said. They don’t know about the therapy. . but you have learnt all these negative ways of looking at things. and that is why I believe I need long term therapy as well. In addition they explained how they thought psychosocial interventions. who understands. Or you need somebody to talk to you and push the right buttons to help sort yourself out.

76 . a few things like that with cognitive therapy. . learning more about oneself and helping to deal with thoughts and feelings. describing it as enabling because it was practical. . he said. The psychosocial interventions that were briefly touched upon were counselling. . being able to return to work. anything that was bothering me. . Counselling Overall people who discussed having counselling were positive about their experiences: The main sort of release point was the counselling. dealing with bereavement issues. which to me was crucial. . happily now saying that at last I’m enjoying life to a greater extent. I think for me it’s about blaming myself . So just things like that. . . you know. . . . and . me understand how I perpetuate the depression . . You know I think they helped quite a bit. Counselling was a positive experience for many because it provided a safe environment in which to talk about their concerns: It was a big relief to have someone who I could tell anything I wanted.Experience of care The benefit of psychosocial interventions to tackle negative thoughts was a prevalent theme. . relaxation therapy and support groups. Cognitive therapy People who had cognitive therapy were positive about it. . that’s what the therapy really helps . self-help material. experiences of different psychosocial interventions are described by people with depression. I’d probably still be severely ill and wouldn’t be. ‘You tend to look back on your day and think of all the failures’ . ‘why don’t you just think of everything that’s been successful?’ So . A particular example was he [therapist] said. thinking that I’m a bad person. focused on the real world and allowed them to begin to help themselves: I could change my thinking and I could thereby change my feeling . it also helped to feel that I was doing something about my problems as well. . In the following sections. when you go lie down to go to sleep. cognitive therapy. . I started doing that . People described how they learnt to change their thoughts to be more constructive and positive: There are things that keep me in a place of being depressed. and not worry about what they might think about it or how it might affect our relationship. If I hadn’t have had the counselling. . Some of the outcomes that people achieved from counselling were: an increase in self-esteem. And you know. and I can expend huge amounts of energy on the mental processes that go into making me responsible for everything that goes wrong in the world.

. Relaxation therapy Two service users described their experience of relaxation therapy: Relaxation therapy . of essentially extremely depressed people talking about suicide. looking after you and some of the things just make me laugh. That’s what I do’. ‘cos it’s everything in the book ties up with other things and you know cognitive therapy for me. It’s really good because it’s all about . people with depression felt relieved to know they were not alone: It was a great source of comfort . In addition. . . Support groups People who had attended support groups were positive about their experiences. . . . some of it is not at all relevant . And to find that in fact you weren’t the only person to feel like that was actually a great relief. . They described these groups as therapeutic because they were able to meet people with similar problems and share their experiences in an environment where there was no stigma. it’s really difficult. to act that way . But it’s a really good one . One read David Burns’ Feeling Good. Another read Dorothy Rowe’s Depression: The Way out of Your Prison: Some of it is relevant. . people who were non-judgemental. when you’re depressed is mighty hard to get started. they won’t say. One described a suicide support group that provided some source of comfort but also had harmful effects: It’s a discussion group of people talking . . pull your socks up. . . which is based on cognitive and behavioural principles: I sat and read this book. And talking about suicidal feelings and suicidal 77 . . . then it’s quite good. but to actually get relaxed when you’re really depressed is damn nigh impossible you know. . and realise that just because you have an opinion or you express yourself a certain way. . You know because it’s so like . . it’s not right or wrong. I’m in there. and you know it’s quite a hefty one. ‘Pull yourself together. . what have you got to be depressed about?’ There is none of that. stop yourself. . It was also a great relief to find . is my chatter box and arguing with it. . A self-help group isn’t group therapy but it is very therapeutic . . ‘That’s me.Experience of care Self-help Two people described using self-help books to cope with their depression. people meeting with a shared interest . It’s very difficult to sort of . . . There are people there who. . . Once you’ve started and got the grasp of it. The mutual support is just unbelievable. to you know.

with taking some tablets. Others who tried medication who did not have positive experiences said they felt that it ‘robbed’ them of feelings. the first tablet . to consider changing your medication. myself.Experience of care methods and yeah. 4. they did not think pills solved the problem or they had a cynical view of drug companies. contentedness It [medication] gave me a feeling that I’ve got some control now of this thing [depression]. It was this lifting again. things are different. c) I didn’t feel comfortable. cutting long gashes in myself. . increased sweating. notably dry mouth. I would say to folk if you feel like you’re not getting any better . And that was the turning point. that they should persevere until they found a drug that works for them: It isn’t a one size fits all . largely because a) I feel that the effects are probably short-term. on the particular medication . Some people were concerned about taking tablets. Many people with depression reported side effects from taking medication. . . . . . from time to time people die on it. A minority also reported experiencing suicidal thoughts as a consequence of their medication: For many years I hadn’t had any suicide thoughts at all. I knew that morning when I woke up that I feel differently. People said that they felt more in control and had greater awareness of the world around them (this was in contrast to other people’s experience of medication): It was exactly 7 weeks to the day that I took . But in a weird perverse way it’s a source of strength and a source of comfort. For those who benefited from a pharmacological intervention. . And I was having some experiences like increased sensitivity to things like noise and colours and feelings. I did start to get sudden images in my head of you know. the majority had positive experiences regarding medication. One person advised that if someone was not benefiting from their current medication. .7 Pharmacological interventions People with depression had mixed views regarding pharmacological interventions. . and I had certainly never thought of cutting myself. but while I was on Seroxat. . However. hair loss. they’re not going to actually resolve the depression. this lifting of overall and just . . go back to your doctor and ask your doctor to change. b) because they do have side-effects and. 78 . .4. they described taking medication as a turning point in their lives. One person described why a pharmacological intervention was not the right treatment for him: I’ve been prescribed antidepressants in the past but I’ve always felt reluctant and apprehensive about taking it. weight gain and problems ejaculating.

and I . which would basically make you go unconscious. . . . and I was talking and laughing. . . that is quite a confusing experience.Experience of care Despite this. nausea. that’s normal’. the nausea was awful. I was just so frightened. but after my experience with the depression I decided I would be prepared to take it . ‘Give me an example’ and I give them an example and they say. I’ve tried to talk about it with the doctors at the hospital and they say. . the most prevalent symptom of which was nausea: Being stupidly pig-headed. but you really don’t . . . short-term and long-term . . but apart from that. the symptoms were so acute it was very frightening. 79 . . just stopped it [Efexor] . the depression again. . that’s not the ECT . including GPs. Only one person reported a positive experience regarding ECT: It all sounds very scary. . so you are asleep.4. the majority had negative experiences because of the frightening nature of the intervention and loss of memory post-treatment: They’d get you to lie down on the bed.9 Healthcare professionals This section covers people’s experience of healthcare professionals. . for the rest of my life if I don’t get it again. But just that 2 minutes when you might have gone into the room and been waiting. . . my mood improved instantly. .4. some people with depression said that the benefits of medication outweighed the potential side effects: You’re given a sheet which tells you what to expect. . I did find that it affected my memory a fair bit. You feel sick. I had no side-effects . I’m very against taking medicine for a long time. I was just completely off my head with depression . And then they give you ECT . . you have a slight headache. And just panic. . if it stops that. I get angry with the professionals that this wasn’t explained that this could happen . 4. really. you don’t see anything because you are anaesthetised. 4. When some people stopped their medication. ‘Oh that’s normal. and give you an anaesthetic in your hand.8 Electroconvulsive therapy Four service users recounted their experience of ECT. they described experiencing discontinuation symptoms. . that’s just normal. and I looked it up on the internet as well. I have massive blanks. nurses and psychiatrists. And you wake up.

against private and NHS. somebody’s actually going to treat me as somebody who has a problem here. it’s the last thing at the end of the day. warm. not as a professional. Psychiatrists People had mixed experience of psychiatrists. it was as though what they [GPs] were saying was. you know. Nurses People said that they did not feel that nurses understood the sensitive nature of their depression. but I really felt that I got no help at all most of the time. helpful and supportive. people were critical of their GPs because they felt that their depression went undetected. Some did not like how psychiatrists tried to illicit information about their childhood experiences. as much time as I want sometimes. I don’t know. She listens and she responds to me as a human being. .’And I know that they’re really busy and I know that they don’t have a lot of time. tender.4. . you didn’t get. However some people had positive experiences of getting a diagnosis of depression and of how their depression was initially managed: I was very low physically and clearly very low mentally. I just don’t think that there is enough. and if you agree I’d like to refer you on to somebody’. and the GP . . kind. there is just not enough funding to be able to . describing the method 80 . . She cares and she’s shown me she cares because she has rung me up before at home and said. .Experience of care GPs As described in Section 4. you know you don’t really understand. ‘Get yourself together’. ‘How are you? Will you come and see me tomorrow?’ because she knows I’m not going to ring and make an appointment because I . ‘Well. . that nurses in the NHS were too busy to talk to their patients and that their attitudes may be because of inadequate training: There’s an awful lot there who . in regards to. actually said. She gives me time. and I’ll be forever thankful for him. you felt as though it was people saying to you. ‘Oh. I mean I’m in isolating mode and things are going wrong. And it was like an immense relief . . . . and. ‘I don’t think I am helping with the right kind of medication for the right reasons. I know better. train the nurses in a certain way. Those with negative experiences described how their GP was lacking in the above characteristics: You just didn’t get listened to. it’s just in your head. which you don’t want. People who had positive experiences of their GPs described them as being sympathetic. for goodness sake pull yourself out of it’. These people felt that they were listened to and responded to: She’s [the GP is] good because she is human.4. .

waiting lists and getting into NHS services were raised. I think because I’d been quite a numerate. I was feeling suicidal. ‘Right. oh . . 81 . way too long. . ‘You’re not depressed. . I was also quite violent at times. . but I knew I had to make a song and dance to get heard. well make it really dramatic.’ I wasn’t pretending exactly. .’ You know. . this isn’t helping me. another described how the psychiatrist prescribed a proper therapeutic dose of antidepressants.4. ‘You are better really’ I was told by one doctor. there was a part of me. continue with the paroxetine. I was really. wet individual. you’re just a very sad lady. saying. The majority had positive experiences: one person described how their psychiatrist was able to change their medication to one with fewer side effects. factual. I swept all the things off his desk you know . One person said that while she was on a waiting list she was unable to cope with her depression: I was referred to the psychiatric hospital for assessment.Experience of care as a ‘text book’ approach that instantly created a barrier. I’ve been on this for eight months. it’s not making me better. . . ‘Oh well. Which again in retrospect was. I mean in my own doctor’s surgery. . way too long. really ill and barely coping. was way. People also had mixed opinions about how their psychiatrist dealt with their medication.’ ‘It takes time.’ 4. Others did not like to discuss feelings in general: I felt my psychiatrist was a very . nothing could ever be pin-pointed or . . ‘Look. organised person. Issues regarding referral. one person felt that she was not listened to when she explained to her psychiatrist that her current medication was not working: He’d [psychiatrist] say something like. Although I think it probably took about two months I believe between the initial sort of GP’s referring letter and getting an appointment. . I just found it annoying. . Again. .’ And if I said. However. . you have to have patience.10 Services The experiences of mental health services were described by people with depression. Another person described how she felt that she had to be violent in her GP’s surgery in order to be referred to NHS services: It’s very difficult to get a hospital bed for quite severe mental illness. Some people said that that they waited too long to be referred to a psychiatrist or receive psychotherapy. . to have someone to talking about feelings and what about this and what about that? And it was . You’ve got to be suicidal . kind of watching what I was doing .

There were routines and I had no responsibilities . One person said that a psychiatric intensive care unit was ‘a place of safety’. which is where she [his wife] was so good. Many times he has forced me to do things and helped me out of the house 82 . Those who had had private treatment had more favourable accounts. In the private hospital you felt like you were being treated as a human being . I was in a place where I didn’t have to think about anything. was the feeling of being institutionalised: In eight weeks. Some felt that their depression had an impact on their children: My sons were very good. There were routines. where they could ‘hand yourself over’ to the care of the service. which isn’t really what you should have to do when you’re growing up. people described a mixture of positive and negative experiences. . Some people said that without their family and carers they would not have been able to cope with their depression: My partner has played a key role in my recovery – he was very supportive during my depression periods – I do not know how I would have coped without him . getting up times. Others described a mental health service as a place where they had no responsibilities. but there was no care . And they would have to make allowances. . Some stated that it was harder for the family and carers than it was for the person who had depression. You felt that yes. I was scared to come out because I was in this enclosed world where I knew what was going to happen. People also had negative experiences of mental health services provided by the NHS. Others described the impact that it had on the partner. . including not feeling cared for. And I won’t knock the NHS because they are obviously very limited to money in a way. often resulting in a change in roles. . and nobody could touch me. . OT [occupational therapy] times. people described how their partners had to take a more active role in daily chores: I found it difficult to relate on the day-to-day things.4. and compared and contrasted the two experiences: The private hospital was.11 Families and carers People with depression described the impact that their condition had on families and carers.Experience of care Once in mental health services. a lot of care in there. For example. . . however. mealtimes. I very quickly became institutionalised myself. She took over those things. Accompanying this. 4. there was a lot of love. . medication times. sincere care. but they missed a lot because of how I was. you could get well here because they cared.

So . self-esteem. I felt I wasn’t letting them down .12 Coping strategies People with depression described coping strategies that they used to overcome their condition. the same pressure is not there. and that . I did decide to just concentrate on small things. Another coping strategy was completing small.’ If you are doing it voluntarily . ‘Look I’m not feeling well. And it was crap at first. Just trying to get through day to day. two things at once. . baking some bread. voluntary work I would definitely advocate because it gives you a sense of . The aim of the review was to explore the experience 83 . I’ve never been in the garden before. to those of a more creative nature such as poetry: Having hobbies. I just felt overwhelmed by it . but gradually it was alright. . . 4. I wasn’t able to do anything about it. you know you start to think. .’ For other people. I believe having a loving and caring partner has helped me get over the most horrible periods of my depression. going for a walk. manageable tasks: When I’m depressed . which ranged from physical activities such as swimming and going to the gym. . this is kind of distracting me a bit.1 REVIEW OF THE QUALITATIVE LITERATURE Introduction A systematic search for published reviews of relevant qualitative studies of people with depression was undertaken. you know. that gets depressed people through because the thing that you can’t think of. I would. 4. I think. . I wanted to do something physical .5. . ‘Yeah. . Distraction was a common coping strategy. . when I was feeling very low. people described how voluntary work helped them to increase their confidence and build up their self-esteem: At the beginning I used to get anxiety attacks and some days I could just phone up and say.4. really.Experience of care in times when I did not feel like doing anything. was how I came out of the suicide attempt. So I started to garden. . . you know pottering around in the garden. . One of the ways in which people distracted themselves from their mental health problem was by having or acquiring a hobby. voluntary work was a coping strategy because the process of helping others allowed them to help themselves. it helps build your confidence. . These strategies were those other than pharmacological and psychological interventions employed by people to manage their depression. In addition. And with my depression. .5 4. . .

in their meta-synthesis of qualitative research in guided self-help in primary care mental health services. 4. HMIC. chronic physical health problems. observational studies People with depression and families/carers None specified of care for people with depression and their families and carers in terms of the broad topics of receiving the diagnosis. Details of the search strings used are in Appendix 8. 2001.4 Themes emerging from the studies Experiencing depression Khan and colleagues (2007). 4. accessing services and having treatment.. childhood events. Elgie.2 Databases searched and inclusion/exclusion criteria Reviews were sought of qualitative studies that used relevant first-hand experiences of people with depression and families/carers.Experience of care Table 5: Databases searched and inclusion/exclusion criteria for clinical evidence Electronic databases Date searched Study design Population Outcomes CINAHL. 2007). Rogers et al.3 Studies considered The search found one systematic review that explored the experience of care for people with depression that met the inclusion/exclusion criteria (Khan et al. Saver et al. 2007) were included in the review that were not already reviewed by Khan and colleagues (2007). 2007).. 4. For more information about the databases searched see Table 5. The GDG did not specify a particular outcome. 2004. found that family conflict. Van Schaik et al. 2005b.. 1997. problems at work. A further seven studies were considered for the review but they did not meet the inclusion criteria (Cooper-Patrick et al. surveys.5. the most common reasons for exclusion were the studies did not report qualitative data or the population did not meet criteria for depression. PsycEXTRA.. PsycBOOKS Database inception to February 2009 Systematic reviews of qualitative studies. Chew-Graham et al.. 2002. the review was concerned with any narrative data that highlighted the experience of care. financial 84 . PsycINFO.. MEDLINE.5.5. Instead. MaGPIe. EMBASE. 2006. Johnston et al. The review team then looked at primary qualitative studies identified by the search and a further two primary studies (Ridge & Ziebland.. 2006.

(3) healthcare professionals seeming unresponsive. These were characterised as: (1) a lack of motivation because of their depression. such as medication. people felt relief because medication helped them cope with difficulties in their day-to-day life. ‘prisoner in my own home’. People taking part in the studies spoke about their depression in terms of the effect on functioning and ability to cope rather than feelings or symptoms. The most common means of expressing their feelings was through metaphor: being ‘on edge’. It has never interested me. Getting a diagnosis of depression For people with depression. while others stated that their GP focused solely on their somatic complaints. Experience of treatment Khan and colleagues (2007) found that taking medication could lead to ambivalent feelings: on the one hand. ‘broken in half’. Accessing help and stigma Khan and colleagues (2007) found that accessing help from primary care could be difficult. There was a feeling among the people in the studies that they were in some way ‘deficient’ because they needed to take antidepressants. In addition. There was also a moral component regarding personal responsibility and the fear of not being able to function in daily life. Feelings 85 . people said that their GP missed opportunities to diagnose their depression. on the other hand. but only on the understanding that it would be for short-term use. people may not have presented their problems in an open manner. ‘boxed in’. they felt a lack of control. ‘a volcano bursting’. seemed uninterested in mental health issues or were purely dismissive of depression when it was suggested. with very little time spent having one-to-one contact with a primary care professional.Experience of care hardship and racism were the most frequent reasons given for causes for depression. Because of feelings of shame and ‘lack of legitimacy’. When the GP or others (family or friends) offered advice to relieve this ambiguity. Some people described their own inability or unwillingness to raise the issue of depression with their GP. People were cautious about telling other people that they were taking medication because of perceived stigma. and (4) a mismatch between how information is offered and how people with depression prefer to seek information. for example: I would never sit down and read something about medicine. (2) stigma associated with depression and/or denial of their diagnosis. There was a possibility that seeking help would ‘threaten an already weakened sense of self’ if treatments were discussed that might be unacceptable to the person. and so on. Saver and colleagues (2007) found that the majority of people received their initial diagnosis from a mental healthcare professional and a minority reported receiving their diagnosis from a GP. Saver and colleagues (2007) described four barriers to accessing help by people with depression. people were more willing to accept medication as a possible treatment. I learned more from watching that commercial on television.

A number of themes were highlighted. Ridge and Ziebland (2006) in their analysis of interview transcripts collected by Health talkonline found that people with deep-seated and complex problems needed longer-term psychological therapy. of letting themselves and others down. I felt rushed along. It should be noted that most of the 86 . in the main. One participant commented that the only ‘option’ was a pharmacological treatment: They just handed me a drug and said go on it right now . None remembered receiving information about the different treatment options such as CBT. had had severe depression.Experience of care of guilt. Saver and colleagues (2007) found that less than half of the people with depression reported receiving information about psychological interventions. problem-solving therapy or IPT. 4. . including feelings of control and helplessness in engaging with treatment. which might influence the success of a self-help intervention for people with depression in primary care. The focus was on the specific components involved in recovery: authenticity. the qualitative analysis and the literature review. Only a minority reported that they had some choice in their treatment options. given a prescription. They saw accessing help as an indication that their personal coping strategies had failed.6 FROM EVIDENCE TO RECOMMENDATIONS This section is a combined summary of themes from the personal accounts. . that getting better meant different things to different people. The main findings of the study were that people needed to understand a language and framework of longer-term recovery to tell their own story of improvement. People said that they used coping strategies such as distraction or thinking of places that were associated with feeling safe and in control. told this will fix it. responsibility and ‘rewriting depression into the self’. and that people needed to assume responsibility for their own recovery. and concerns about long-term changes to their personality were also expressed. to explore the approaches and meanings attributed to overcoming depression. Recovery Ridge and Ziebland (2006) analysed the interview transcripts (collected by Healthtalkonline) of 38 men and women who. The majority of the interviewees had used and valued talking therapies as a means of gaining insight into their thoughts and feelings. Self-help and other coping strategies Khan and colleagues (2007) synthesised qualitative studies of patient experiences of depression management in primary care to develop a framework for a guided self-help intervention with the aim of providing a potential solution to the problem of the gap between demand for CBT and supply of trained therapists. Recovery involved the need to understand the ‘authentic self’.

It was felt that an emergency number to call would be a lifeline for people who live alone and have no carer support. Feelings of 87 . family and work associates. Two people in the personal accounts (B and E) found it difficult to access support when needed. Therefore written information is crucial. such as via video or DVD. One person (G). There was also a concern that when a person is severely depressed they may find it difficult to concentrate on what is being said. External stigma was felt from employers and colleagues. had empowered himself through the internet and had built up a wide network of fellow sufferers. 4.1 Understanding depression Both the personal accounts and the literature reveal that lack of information from professionals is a barrier to coming to a full understanding of depression.6. The literature also revealed that accessing help may be a problem for some people first experiencing symptoms because of stigma associated with having a mental health problem (see Section 4. Lack of clarity about how care is organised may increase the person’s distress. Therefore.Experience of care personal accounts received were from people who either have or have had severe and/or chronic depression. One person (B) said that it would be helpful if professionals could be clear about the purpose of any appointments offered. as evidenced by personal account C. whether it was during the initial stages of being diagnosed or after years of having treatment.6. the qualitative analysis and the literature. Such means of support would be particularly helpful for people with long-term. the range of treatments available and the role of the mental health team.3). This was experienced both externally and internally. Despite these limitations.2 Accessing help and getting a diagnosis of depression Accessing help was also a prevalent theme in the personal accounts. despite having had depression for some years.3 Stigma Stigma was frequently discussed in the personal accounts.6. which may leave them unmotivated to raise the issue of depression with their GP. the qualitative analysis and in the literature.6. a number of themes were identified that were present in all three sources of evidence. although it should be recognised that people with mental health problems may respond to information provided in other forms. Lack of accessible information is a particular issue for people from black and Asian minority ethnic groups. severe depression. it is acknowledged that the themes that run through the personal accounts may not be applicable to people who have other forms of depression. who had been given no information. 4. 4. but many also felt internal stigma and kept their depression concealed from friends.

In the personal accounts. Most of the personal accounts spoke of the importance of a relationship with professionals that was non-judgemental and supportive. another said her psychiatrist did not listen when she said her medication was not working. F and G) expressed negative views. In the personal accounts. such as the psychiatrist being unsupportive and cursory in their attention.4 Recognising depression Recognition of depression and the severity of symptoms was also a prominent theme in the three forms of evidence. 4. But as one person (B) pointed out. In the personal accounts. the relationship with the GP was seen negatively if the GP failed to recognise depressive symptoms or focused solely on the person’s somatic symptoms. her self-report was not listened to closely enough and the severity of her depression was underestimated. with lack of understanding about depression being cited as a common complaint. In the qualitative analysis there were mixed views about psychiatrists.6. The main area of criticism concerned the quality of contact with the GP (see Khan et al. People who had positive experiences highlighted the sympathetic.6. particularly in the way that they prescribed medication. sometimes being well-meaning and supportive is not enough.5 Relationships with healthcare professionals The relationship with the GP was a prevalent theme in the personal accounts. One person (B) felt that her diagnosis should have been made by a qualified and experienced professional. supportive and helpful qualities of the GP. the qualitative analysis and the literature.Experience of care shame were expressed and also an anxiety that asking for help would lead to being offered interventions that they did not want. most found their GPs helpful and understanding. The relationship with nurses was not as positive in both the personal accounts (see B) and the qualitative analysis. She felt that while her primary care practitioners and counsellors were pleasant and accommodating. Some people felt that their psychiatrist was able to work with them to find the right medication and the correct dose. such as medication (the person in account D said that the idea of taking tablets accentuated the feeling of being mentally unwell). some people had neutral views about their psychiatrist while three people (C. 2007) – a short appointment when a person is distressed is not long enough and people with depression are unlikely to ask for a longer appointment. In the literature and qualitative analysis. two people (B and G) commented that they felt that the severity of their depression was not properly recognised within primary care.. 4. In the qualitative analysis and the literature. people spoke about how depression is often not recognised and that physical problems may mask the depressive symptoms or may not be seen as part of the depressive symptomatology. A number of people commented that the relationship 88 .

Khan and colleagues (2007) found that family conflict and childhood events were among the most frequent reasons given for causes for depression. It emerged from the qualitative analysis that some people ascribed the onset of their depression to certain life events.6. and that ideally there should be some choice in terms of the gender of the therapist and their therapeutic approach. for some. physically or emotionally. people with depression described some of the negative thoughts that they had experienced and some described suicidal thoughts and behaviour. Howe (1995) explains that: Internal psychological states and our ability to cope with the external demands of life have roots which reach right back into childhood. those who had private treatment had.Experience of care between patient and therapist is of prime importance.7 Experience of depression and its possible causes In both the personal accounts and the qualitative analysis. they also used metaphor and allusion to explain their symptoms. In both sources of evidence. The robustness of our early internal representations of self and others lays down the pattern of our future psychological strengths and weaknesses. 4. views were mixed. 4. Early attachment relationships that were lost or broken leave people feeling that they cannot control the important things in their lives. was a positive outcome. For the person in account D. abuse or conflict of one form or another and many of them linked this directly with the onset of their depression. being a carer of someone with schizophrenia meant that he had to hide his symptoms of depression to fulfil his role as a carer. Once in mental health services. complex problems in childhood were compounded by multiple difficulties in adulthood.6 Experience of services Both the personal accounts and the qualitative analysis described experiences of mental health services. In the qualitative analysis some people said that they were able to experience life differently since being depressed which. they are not able to control what happens to them. Attachment relationships in which sexual or physical abuse took place often leave the individual with feelings of passivity and worthlessness. Two people (A and B) commented that it is often seen as the patient’s ‘fault’ if they do not benefit from psychological treatment. Many people said that they waited too long to be referred to a psychiatrist or receive psychological treatment. including childhood experiences. The majority of the personal accounts also reported childhood events such as trauma. Without support they remain 89 . a feeling of fatalism and helplessness sets in. When children feel that no matter what they think. when the counsellor or therapist should take some responsibility for a lack of therapeutic effect.6. For many people. say or do. on the whole. more positive experiences.

For those who feel hopeless and helpless. it cannot be recommended as a course of treatment in this guideline (see Chapter 8). such as CBT and IPT. depression will be a major factor in the person’s life. while it was good to meet other people. The personal accounts express mixed views about support groups: one person (D) was very positive about them. She felt that for counselling to be effective. The study by Ridge and Ziebland (2006) confirms the opinions of the topic group and the testimony from the personal accounts that people with ‘deep and complex problems felt the need for longer term therapy’. that is. Overall. people in the qualitative analysis were positive about counselling. but less positive about relaxation therapy because people with severe depression find it difficult to relax. people expressed a need for psychosocial interventions when they attributed the cause of their depression to psychological processes rather than a ‘chemical imbalance’ and to help them cope with negative thoughts. as were people in the personal accounts. 90 . Psychosocial interventions This was a theme of both the personal accounts and the qualitative analysis. Another felt that the counselling she received was unsatisfactory: she was asked inappropriate questions. The study by Saver and colleagues (2007) points to the fact that few people received information about psychological therapy and the different treatments. Those that have had long-term psychodynamic therapy report that it has been helpful in their understanding of themselves and their depression and that until they have worked through and repaired the damage experienced in childhood. although concerns were raised by two people (B and E). the counsellor needed to both listen and question skilfully. 4.Experience of care emotionally vulnerable to setbacks and upsets.6. One found counselling inadequate for her needs because it did not get to the ‘root’ of her depression and indeed did not stop her depression from becoming more severe. self-help books and support groups. The service user and carer topic group do acknowledge. that psychological treatment offered by the NHS in the form of CBT does not go far enough in addressing the trauma experienced in childhood. that as there has been little research into the efficacy of long-term psychodynamic therapy. The view of relaxation therapy is borne out in personal account B. incorrect assumptions were made about her life. In the qualitative analysis. and she felt that she did not talk enough during the sessions. she gained no therapeutic value from attending. but another (E) said that. however.8 Experiences of treatments Psychological therapy There was a strong feeling within the service user and carer topic group that the excerpt from Howe (1995) in the section above highlights the reasons why many people opt for private therapy. people were generally positive about cognitive therapy. In the qualitative analysis. depression is often the psychological result.

Several personal accounts spoke of difficulties in getting paid employment: one person (C) stated that both their college and job centre could not help until their condition was stable. was unable to work and had no income.6. 4. Others benefited from it.9 Coping strategies It is evident from the personal accounts and the literature review that people who have had depression for a long time develop positive coping mechanisms that enable them to manage their illness.Experience of care Khan and colleagues (2007) synthesised qualitative studies of patient experiences of depression management in primary care to develop a framework for a guided self-help intervention. In the literature. These mechanisms range from exercise (A) or personal faith (C). In personal account G. including loss of memory after treatment. they felt a lack of control. To contextualise this theme. Some people did not find antidepressants helpful. Only one person had a positive experience with no side effects. Medication There were mixed reports regarding medication. having a hobby. the person had only worked in paid employment 91 . The qualitative analysis also identified a number of coping strategies such as distraction.6. to readjusting one’s life to be able to manage depression. activities and voluntary work. The majority of people who had ECT had negative experiences. 4. some of the literature regarding this topic that was not identified in the systematic search is briefly described below. others were concerned about taking tablets. particularly in the form of a ‘drug cocktail’. and another (B) was self-employed when she became ill. one person (B) felt strongly that getting the appropriate medication promptly is vital and that there should be intense support before the antidepressive effects are experienced. Electroconvulsive therapy This theme was only present in the qualitative analysis. one person (A) commented on the weight gain associated with the medication leading to self-esteem issues and feeling more depressed. it emerged that taking medication could lead to ambivalent feelings: on the one hand. people felt relief because medication helped them cope with difficulties in their day-to-day life.10 Employment The theme of employment was only present in the personal accounts. on the other. The majority of people in the qualitative analysis said that antidepressants were beneficial. In the personal accounts. despite some experiencing side effects. From the personal accounts there are issues for those with long-standing depression when it comes to accessing and remaining in employment.

that for people with depression a return to work or continuing with work can aid the recovery process. The Health and Safety Executive (2008) reported that in 2006/07. One person (A) spoke of colleagues not being keen for her to return to work. the carer has built her career around self-employment so that she has time to care. The results showed that. people with mental health problems may turn up for work even if they are feeling unwell rather than be labelled as mentally ill by their employers and co-workers. Fearing possible stigma or discrimination. Other personal accounts spoke of experiences in work. A report by Waddell and Burton (2006) concluded that work was generally beneficial for both physical and mental health and well-being. Crowther and colleagues (2001) found that participants who received supported employment were more likely to be in competitive employment than those who received prevocational training (34% compared with 12% at 12 months). Another person (E) expressed the fear of getting too ill to work. but is also able to maintain a life outside caring. they may remain absent from their place of employment or unemployed for considerable periods of time. but was doing voluntary work with mental health and disability organisations. depression or anxiety that was caused or exacerbated by their current or past employment.Experience of care for 8 months in the 8 years he had had depression. A report by the Royal College of Psychiatrists (2008) found two studies that analysed employment schemes in people with mental health problems. conversely. Overall. Rinaldi and colleagues (2008) examined a supported employment scheme run by South West London and St George’s Mental Health NHS Trust. In a systematic review of 11 RCTs comparing prevocational training or supported employment for people with severe mental illness with each other or with standard community care. following 92 . and instead of returning to her normal activities she was marginalised from external meetings and confined to certain tasks.000 people in the UK reported they were experiencing stress. It was estimated that 13. Clinical research and government reports suggest that employment plays a part both in exacerbating stress leading to depression. and should offer the individual some influence over how the work is done and a sense of self-worth. The Sainsbury Centre for Mental Health (2007) also identified the loss in productivity that occurs when employees come to work but function at less than full capacity because of ill health (termed ‘presenteeism’). the beneficial effects of work were shown to outweigh the risks and to be much greater than the harmful effects of longterm unemployment or prolonged absence because of sickness. Once people with depression become too ill to work. that it can be crucial component in aiding the recovery process. she also had the support of her manager in whom she confided and who helped with work pressures. It advised that the type of employment should be healthy and safe. an estimated 530. but also. however. some people commented that stressful situations at work contributed to the onset of their depression. However. depression or anxiety.8 million working days (full-day equivalent) were lost in 2006/07 through work-related stress. In the qualitative analysis. The anecdotal evidence from the personal accounts suggests. but with the help of her GP did not have to say that the occasional day or week off with illness was because of depression. The issue of employment is also important to carers: in personal account H.

Caring can radically change the relationship between partners and between parents and children. 2008. 2009) provide guidance on how this can be achieved. to perhaps currently living with dysthymia. one from an adult caring for her partner (H) and one from a teenage boy caring for his mother (I). the person is able to assume responsibility for their illness through gaining insight into themselves. and their place in the world. The term ‘recovery’ was the cause of significant debate in the service user and carer topic group and had different meanings for different people. The needs of young carers should be recognised and addressed and recent publications from the Social Care Institute for Excellence and the Department of Health (Department of Health et al. The carer in account H felt more like a mother than a partner and the young carer (I) said that he became an adult when he cared for his mother. Roberts et al. But several themes did emerge. For others the term ‘recovery’ was important in demonstrating the positive shift from being severely depressed with an inability to ‘function normally’. Treatments and professionals were seen as the ‘tools’ needed to aid recovery. For some it meant an absence of depressive symptoms and an ability to function fully to one’s potential.12 Families and carers The literature search did not identify studies of carer experience and the two personal accounts offer very different perspectives. Greene et al.. They explain that as the process of recovery develops. with just a few residual symptoms that are manageable. The report by 93 . But for other long-term sufferers. their concept of themselves and others around them. Both carers also spoke of the different aspects of caring: undertaking practical tasks for the person. Department of Health et al. 4. Both carers reported that having interests that took them away from caring for a few hours was extremely important..6. where the user is able to live a full and productive life. It should be recognised that young carers might marginalise themselves from their peer group and experience other social and educational disadvantage..Experience of care the integration of employment specialists into CMHTs.11 Recovery In the study by Ridge and Ziebland (2006). their thought processes. but that she became a ‘normal bossy Mum’ again when she was well. ‘recovery’ was a term that they would not use (‘selfmanagement’ being perhaps a more appropriate term). The conclusion drawn supports the use of individual placement specialists in clinical practice in CMHTs. 2008. 4. the term ‘recovery’ is used to describe the process by which people learn to understand and then manage their illness..6. there was a significant increase in the number of clients with various diagnoses (31% with depression – unspecified severity) engaged in mainstream work or educational activity at both 6 and 12 months. 2008. The personal accounts both conveyed the experience that caring is rewarding but challenging. and offering emotional support.

although some people had to cope with their depression alone. 94 . This is especially important when a person has severe depression or is subject to the Mental Health Act7.7. and obtaining informed consent When working with people with depression and their families or carers: ● build a trusting relationship and work in an open.7.7.1.2 4. someone who is not known to the person with depression) if needed5.1.Experience of care Roberts and colleagues (2008) suggests that the needs of young carers could be more effectively addressed by respecting their anxieties and acknowledging their input and skills. 4. with some people stating that depression was harder for family members and carers than for themselves.3 4. Many people also commented on the supportive nature of family members and carers. Some people remarked on the change of roles that occurred as a result of one person having depression.7. support groups and other local and national resources6. 7Ibid.1. explaining the different courses of depression and that recovery is possible ● be aware that stigma and discrimination can be associated with a diagnosis of depression ● ensure that discussions take place in settings in which confidentiality. Any wording changes have been made for clarification only. privacy and dignity are respected. It is also recommended that young carers should be included in their family member’s care planning.7.1 RECOMMENDATIONS Providing information and support. Make all efforts necessary to ensure that a person with depression can give meaningful and informed consent before treatment starts.1.7 4. The impact of depression on families and carers was a prolific theme in both the personal accounts and the qualitative analysis.4 5The evidence for this recommendation has not been updated since the previous guideline. 4. When working with people with depression and their families or carers: ● provide information appropriate to their level of understanding about the nature of depression and the range of treatments available ● avoid clinical language without adequate explanation ● ensure that comprehensive written information is available in the appropriate language and in audio format if possible ● provide and work proficiently with independent interpreters (that is. Inform people with depression about self-help groups. engaging and non-judgemental manner ● explore treatment options in an atmosphere of hope and optimism. 6Ibid.1 4.

consider developing advance decisions and advance statements collaboratively with the person.2. ethnic and religious backgrounds when working with people with depression.1. if the person agrees. Ensure competence in: ● culturally sensitive assessment ● using different explanatory models of depression 8Depression is described as ‘chronic’ if symptoms have been present more or less continuously for 2 years or more.7. consider: ● providing written and verbal information on depression and its management. and sensitive to. and helping families or carers to access these ● negotiating between the person and their family or carer about confidentiality and the sharing of information.5 Ensure that consent to treatment is based on the provision of clear information (which should also be available in written form) about the intervention. Give copies to the person and to their family or carer.7.1 Supporting families and carers When families or carers are involved in supporting a person with severe or chronic8 depression.4.7. Record the decisions and statements and include copies in the person’s care plan in primary and secondary care.1 Advance decisions and statements For people with recurrent severe depression or depression with psychotic symptoms and for those who have been treated under the Mental Health Act.Experience of care 4. 4.7.7. and be aware of the possible variations in the presentation of depression. physical and mental health needs if necessary ● providing information about local family or carer support groups and voluntary organisations. covering: ● what it comprises ● what is expected of the person while having it ● likely outcomes (including any side effects).4 4. diverse cultural.3. including how families or carers can support the person ● offering a carer’s assessment of their caring.7. 4. 4.2 4. 95 .1 Working with people from diverse ethnic and cultural backgrounds Be respectful of.3 4.7.

Consider providing all interventions in the preferred language of the person with depression where possible9. ● 4. 9The 96 .7.Experience of care addressing cultural and ethnic differences when developing and implementing treatment plans ● working with families from diverse ethnic and cultural backgrounds.2 evidence for this recommendation has not been updated since the previous guideline.4. Any wording changes have been made for clarification only.

that many cases go unrecognised (Del Piccolo et al.. which has contributed to increased access to low-intensity psychosocial interventions. 2009). in the UK these developments have included the introduction of graduate mental health workers (Department of Health. The updated reviews for guided self-help. 1992. including computerised CBT (NICE. that will bring about the single biggest change in the provision of effective treatments for depression in primary and secondary care. for example diabetes (Von Korff et al. that go undetected and untreated (Thompson et al.Case identification and service delivery 5 CASE IDENTIFICATION AND SERVICE DELIVERY 5. 10For this guideline update. Donoghue & Tylee. 2005).. 1994). 1998. Kessler et al. enhanced care and integrated care (now re-named enhanced care). Raine et al.. however. Where depression is recognised. which are more likely to recover spontaneously. computerised CBT and exercise (now termed physical activity programmes) have been moved to Chapter 7. 2001. 2001). 97 . These responses have included developments in the treatment of depression in primary and secondary care. and the updated review for ECT can be found in Chapter 12. stepped care.400 new psychological therapists. The sections from the previous guideline on screening (now re-named case identification). the organisational and professional structures of primary and secondary care mental health services. 2002.1 INTRODUCTION10 The starting point for providing effective treatment for depression is the recognition of the problem and the first point of access is usually primary care. organisational developments such as collaborative care. There is evidence. Von Korff & Goldberg. In addition to efforts to improve recognition of depression. 2003). It is this later development. however. with the majority of people continuing to be managed in primary care. with £340 million of funding over 6 years along with 3.. care often falls short of optimal recommended practice (Katon et al. 2003).. 2007. Since the publication of the previous guideline in 2004. 1996) and outcomes are correspondingly below what is possible (Rost et al. 1997.. 2000). The concept of ‘stepped care’ advocated in the previous guideline has been embraced by many commissioners and providers in the NHS and is now being taken forward by the Improving Access to Psychological Therapies (IAPT) programme (Department of Health. suggest that clinically significant depression (moderate to severe depressive illness) is detected by GPs at later consultations by virtue of the longitudinal patient–doctor relationship and it is milder forms. a number of responses have been developed over the past 20 or so years to address the problem of suboptimal treatment. IAPT. This is a cause of considerable concern.. all sections of the ‘Service-level and other interventions’ chapter in the previous guideline were reviewed. More recent studies. and the development and adaptation of models for the management of chronic medical conditions. non-statutory support and crisis resolution and home treatment teams remain in this chapter. NICE.

. 1992. 5. Kessler et al. this under-recognition of depression may be focused more on mild depression than on moderate or severe depression (Kessler et al. go undetected and untreated (Thompson et al. some studies suggest that clinically important depression (moderate to severe depressive illness) is detected by GPs at later consultations by virtue of the longitudinal patient–doctor relationship and that its milder forms. the development of new and existing staff roles in primary care and the introduction of mental health specialists into primary care. Mitchell and colleagues (2009) suggest that GPs are able to rule out depression in most people who are not depressed with reasonable accuracy but may have difficulty diagnosing depression in all true cases. including physical health problems. 2000.1 As stated above the accurate identification of depression is an essential first step in the management of people with depression. 2002).. The identification of depression in adults with a chronic physical health problem is covered in a related NICE guideline (NICE.. This guideline focuses on identifying depression in primary care and community settings. These approaches to service delivery fall into three main groups. 5. 2002). However. as noted below. As stated above.. Other authors draw attention to the dangers of the erroneous diagnosis of depression in patients with a slight psychological malaise and few functional consequences that can lead to the risk of unnecessary and potentially dangerous medicalisation of distress 98 . including systematic approaches for organising care and making available appropriate treatment choices..2 THE IDENTIFICATION OF DEPRESSION IN PRIMARY CARE AND COMMUNITY SETTINGS Introduction 5. This includes both people who have sought treatment because of depressive symptoms and those being treated for other conditions. even this suggests that non-clinically important depression may go undetected initially. 2009c). a view which is supported by a recent meta-analysis of 37 studies of GPs’ unassisted ability to detect depression (Mitchell et al. Studies indicate that up to 50% of people with depression are not recognised when they attend primary care (Williams et al.Case identification and service delivery This chapter focuses on two main issues: the identification of depression in primary and secondary care and the range of different service delivery mechanisms that have emerged in recent years.2. 2003). it has been suggested that GPs fail to accurately diagnose depression (Goldberg & Huxley. 1995).. However. 2005). although almost half of the people prescribed antidepressants were not depressed (Kendrick et al.2. Kessler et al. 2009).2 Identifying depression – a primary care perspective For over 40 years.. More recent studies suggest that the probability of prescribing antidepressants in primary care is associated with the severity of the depression. which may recover spontaneously.

5. 2006). differential reporting of symptoms may lead to depression being under-diagnosed in men.3 Factors related to the person with depression People may have difficulty in presenting their distress and discussing their concerns with their doctor.. Although depression is more frequent in women. A number of other factors may also influence the identification of depression. and also to clarify the accuracy of GPs in diagnosing anxiety disorders. 2004. Given the modest prevalence of depression in most primary care settings the number of false positive errors (people who are incorrectly identified as being at risk of depression) is larger than the number of false negatives (those falsely identified as not being at risk of developing depression). explain low rates of help-seeking among young adults. in fact. In deprived areas. or that such symptoms should not be discussed at all. it has been suggested that contact with primary care may be of little significance when set against the magnitude of their other problems (Rogers et al. at least in part. primary care physicians have been shown 99 . 1996).. in particular. including those with severe distress (Biddle et al.. or that it is not possible for doctors to listen to them and understand how they feel (Pollock & Grime. 2006). May et al.. From the perspective of the person with depression. 2001).2. Further work is clearly needed to examine the subsequent outcome of those false positive and false negative diagnoses.4 Practitioner factors The construction of ‘depression’ as a clinical condition is contested amongst GPs (Chew-Graham et al. 2003).. The MaGPIe Research Group (2005a. Pilgrim & Dowrick. especially when they are uncertain that depression is a legitimate reason for seeing the doctor (Gask et al. Reasons for lack of recognition fall into four themes: factors related to the person with depression. People may have beliefs that prevent them from seeking help for depression such as a fear of stigmatisation. 2003). 2002. or that antidepressant medication is addictive or they may misattribute symptoms of depression for ‘old age’. 2006). 5. 2006). may complain less of depressed mood and instead somatise their depressive symptoms (Rabins.Case identification and service delivery (Aragones et al. 2000. adjustment disorders and broadly defined distress. and that GPs are.. Gask et al.. They may be wary of opening a ‘Pandora’s box’ in time-limited consultations and instead collude with the person with depression in what has been called ‘therapeutic nihilism’ (Burroughs et al. Physical comorbidity can also make the interpretation of depressive symptoms difficult. effective at identifying mental health problems in patients they know. organisational and societal factors. ill health or grief. however some people believe that the GP is not the right person to talk to. and practitioner. Older adults. The person with depression may feel that they do not deserve to take up the doctor’s time. Negative perceptions about the value of consulting a GP for mental distress may. 2005b) suggests that the relationship is important.2..

closer monitoring or even preventative action.Case identification and service delivery to view depression as a normal response to difficult circumstances. 1998). Pilgrim & Bentall. It is therefore important that the healthcare professional recognises and accepts their own reaction to people presenting with depression so that they can acknowledge and go on to diagnose depression.2..5 Organisational factors The trend in the UK for mental health services to be separate from mainstream medical services may disadvantage people with depression who may have difficulties in attending different sites and/or services for mental and physical disorders. 2004). Screening programmes detect people at risk of having the condition or at risk of developing the condition in the future. 1989). together with limited access to mental health services. Chew-Graham et al. They do not establish a diagnosis but give some indication of any action that may be required. 1996. 5. management or treatment.2. The oft-described barrier of stigma has to be set against the arguments that depression is a social construction within which chronic distress or unhappiness are medicalised (Ellis. Screening has been defined as the systematic application of a test or enquiry to identify individuals at high risk of developing a specific disorder who may benefit from further investigation or preventative action (Peckham & Dezateux. Those falsely 100 . add to professionals’ reluctance to encourage patients to disclose their distress (Popay et al. Primary care practitioners may also lack the necessary consultation skills or confidence to correctly diagnose late-life depression. illnesses or life events (May et al. and depression may be under-diagnosed because of dissatisfaction with the types of treatment that can be offered.7 Shifting the emphasis from screening to identification The identification of people with a disease is often referred to as screening (and was the term used in the previous guideline). people who are incorrectly identified as being at risk of developing a condition (false positives) can be subject to further possibly intrusive.6 Societal factors The barriers described are likely to be particularly difficult for the economically poor and minority populations who tend to have more health problems and are more disabled. harmful or inappropriate investigations. 5. and then discuss a range of possible interventions.2. such as further diagnostic investigation. 2007. 1999) and the suggestion that chronic unhappiness is not ‘treatable’ in the normal curative or therapeutic sense. Screening is not necessarily a benign process (Marteau.. Since screening tools are never 100% accurate. especially a lack of availability of psychological interventions. 5. 2008). Organisational factors that inhibit the identification and disclosure of symptoms and problems..

significant physical illnesses causing disability.2. 2005]) may improve the specificity of the screening questions. Within the NHS.. Others. biomedical approach. It has been suggested that using an additional question (‘is this something with which you would like help?’ [Arroll et al. the expenditure of resources on patients who may gain little benefit (many patients who are detected by such an approach may be mildly depressed and recover with no formal intervention). Critics of routine screening for depression have advanced a number of arguments against it. the lack of empirical evidence for benefit to patients. 1996). diverting them from a broader bio-psychosocial approach to both diagnosing and managing depression (Dowrick. feedback to patients or case management) is considered below and the GDG felt it important to first address the value of case identification systems alone. such as those with a family history of the condition or chronic physical health problems with associated functional impairment. 2008). caution that the use of such screening instruments may encourage practitioners to take a reductionist. 5. such as dementia. before going on to consider the benefits of integrated systems. 2004). Others (for example. Macmillan et al. 2002. These include the low positive predictive value of the instruments (that is.Case identification and service delivery identified as not being at risk of developing a condition (false negatives) will also suffer by not being given the further investigation they need. 2005) have. diabetes) is now part of routine clinical work for GPs as stipulated by the GMS Contract (Ellis. The effectiveness of such interventions (for example. Palmer and Coyne (2003) also go on to make a number of suggestions for improving recognition. however. Pignone et al. These issues are well covered by Palmer and Coyne (2003) in their review of screening for depression in medical settings. however. in addition to other NICE mental health guidelines. Evidence. case identification of depression in people with some chronic conditions (for example.8 Case identification Introduction The previous NICE guideline on depression. NICE. focusing on patients with previous histories of depression and people known to have a high risk of developing depression. suggests that such ultra-short screening instruments may fail to detect depression (Mallen & Peat. however. but it should be noted that the systematic review of interventions conducted in support of the recommendations by these groups have included the need for follow-up interventions. 101 . and the diversion of resource away from more seriously depressed and known patients who may be inadequately treated as a result. considered the case for general population screening for a number of mental health disorders and concluded that it should only be undertaken for specific high-risk populations where benefits outweigh the risks (for example. These were people with a history of depression.. recommended the use of screening of depression for the general adult population. including ensuring effective interventions for those identified. many patients who screen positive do not have depression). or other mental health problems. 2004b)..

This would lead to an under-estimation of the prevalence of the disorder. The identification tools were assessed in consultation (which included primary care and general medical services) and community populations. 1999). Definition Case identification instruments were defined in the review as validated psychometric measures that were used to identify people with depression. 1998.Case identification and service delivery A history of depression has been identified as a significant factor in future episodes. For example.. Patient Health Questionnaire (PHQ). Summary statistics used to evaluate identification instruments Sensitivity. The sensitivity of an instrument refers to the proportion of those with the condition who test positive. It should be noted that depression can frequently be comorbid with other mental health problems. 102 . having a history of depression produced a positive predictive value (see below) roughly equal to that produced by using a depression case-finding instrument (Centre of Epidemiology Studies-Depression – CES-D) (0.or two-item measures. Zanarini et al. Studies that did not clearly state the comparator to be DSM–IV or ICD–10. positive predictive validity and negative predictive validity The terms ‘sensitivity’ and ‘specificity’ are used in relation to identification methods discussed in this chapter. The following sections review available case identification instruments. and dementia (Ballard et al. Hospital Anxiety and Depression Scale (HADS).. contribute to inadequate care and make for poor planning and costing of the need for treatment. Skodol et al. Geriatric Depression Scale (GDS)..28). As the specificity of an instrument increases. This is important so that healthy people are not offered treatments they do not need.25 compared with 0. the number of false negatives it detects will decrease. As the sensitivity of an instrument increases. The review was limited to identification tools likely to be used in UK clinical practice. In fact. including borderline personality disorder (for example. the number of false positives will decrease. An instrument that detects a low percentage of cases will not be very helpful in determining the numbers of patients who should receive a known effective treatment. that is. Studies were sought that compared case identification with one of the above instruments with diagnosis of depression based on DSM–IV or ICD–10 criteria. 1999). 1996). ‘Gold standard’ diagnoses were defined as DSM–IV or ICD–10 diagnosis of depression. the Beck Depression Inventory (BDI). The specificity of an instrument refers to the proportion of those who do not have the condition and test negative. Centre of Epidemiology Studies-Depression (CES-D). a study of 425 primary care patients found that 85% of those who were depressed had had at least one previous episode (Coyne et al. specificity. General Health Questionnaire (GHQ). used a scale with greater than 28 items. Zung Self Rated Depression Scale and any one. or did not provide sufficient data to be extracted in the meta-analysis were excluded. as many individuals who should receive the treatment will not do so. This suggests that careful assessment of relevant instruments is required if a number currently in use appears to have no more predictive value than a history of depression..

the area under the curve (AUC) measures how close the tool gets to the theoretical ideal. 103 . The example above illustrates some of the main differences between positive predictive values and negative predictive values in comparison with sensitivity and specificity. that is. While this is never achieved in practice. prevalence explicitly forms part of their calculation (see Altman & Bland. 0.9 as ‘excellent’. and less than 0. Therefore although these statistics are concerned with issues probably more directly applicable to clinical practice (for example.Case identification and service delivery To illustrate this. the GDG defined values above 0. For example. a positive test result is correct in only 40% of cases.3 as ‘poor’.5 is better than chance. in this example.000 people are given a test which has 90% sensitivity and 85% specificity. 1994a). 1994a). When describing the sensitivity and specificity of the different instruments. The positive predictive value of the test (the number correctly identified as having depression as a proportion of positive tests) is 40% (90/90 135). theoretically these estimates are not affected by prevalence. but classifies 135 incorrectly as having depression (false positives). As discussed above. 1.5 as ‘low’. and the negative predictive value (the number correctly identified as not having depression as a proportion of negative tests) is 98% (765/765 10). The advantage of this approach is that sensitivity and specificity can be applied across populations (Altman & Bland. but the test detects only 90 (true positives). 1994b).5 to 0. which plots sensitivity (expressed as a per cent) against (100specificity) (Figure 4).7 as ‘moderate’. It is also known that 900 people do not have depression. it would have 100% specificity and pick up all true positives with no false positives. A perfect test would have an AUC of 1. and a test with AUC above 0.8 to 0. Receiver operator characteristic curves The qualities of a particular tool are summarised in a receiver operator characteristic (ROC) curve. 0. the main disadvantage is that clinicians tend to find such estimates more difficult to interpret. the probability that a person with a positive test result actually has depression). 1994b). When the prevalence of a disorder is low in a population this is generally associated with a higher negative predictive value and a lower positive predictive value.9 as ‘good’. Therefore the higher false positives often associated with samples of low prevalence will not affect such estimates. and the test correctly identifies 765 of these (true negatives). they are largely dependent on the characteristics of the population sampled and cannot be universally applied (Altman & Bland. It is known that 100 people in this population have depression. from a population in which the point prevalence rate of depression is 10% (that is. because these measures are based on sensitivity and 100-specificity. while a negative result can be relied upon in 98% of cases. On the other hand. leaving 10 undetected (false negatives).3 to 0. A test with perfect discrimination would have an ROC curve that passed through the top left hand corner. 0. 10% of the population has depression at any one time). sensitivity and specificity do not necessarily depend on prevalence of depression (Altman & Bland. sensitivity is concerned with the performance of an identification test conditional on a person having depression. However. For both positive and negative predictive values. Therefore.

Case identification and service delivery Figure 4: Receiver operator characteristic curve Negative and positive likelihood ratios Negative (LR ) and positive (LR ) likelihood ratios are thought not to be dependent on prevalence. 2003). Of these studies.. Diagnostic odds ratios The diagnostic odds ratio is LR /LR . 2009c). A value of LR 5 and LR 0. 2010). LR is calculated by sensitivity/1-specificity and LR is 1-sensitivity/ specificity. Studies considered A total of 126 studies met the eligibility criteria of the review. This was undertaken as a joint review for this guideline and the guideline for depression in adults with a chronic physical health problem (NICE. 2003). Depression in Adults with a Chronic Physical Health Problem (NCCMH. Details of the search strings used are in Appendix 8. 45 were on people with chronic physical health problems11 and 50 were on older people (over 65 years of age). 16 11Data for the population with chronic physical health problems and information about the included studies is presented in the related guideline. Information about the databases searched and the inclusion/exclusion criteria used can be found in Table 6. a value of 20 or greater suggests a good level of accuracy (Fischer et al. Databases searched and inclusion/exclusion criteria The review team conducted a new systematic search for cross-sectional studies to assess tools for identifying depression..3 suggests the test is relatively accurate (Fischer et al. 54 studies were conducted in consultation samples. 104 .

2008). The most common reason for exclusion was a lack of a gold standard (DSM/ICD) comparator (see Appendix 20 for further details). Patient Health Questionnaire The PHQ developed out of the more detailed Primary Care Evaluation of Mental Disorders (PRIME-MD) (Spitzer et al. nine on the BDI – short form. five on the PHQ-2. PHQ. 16 on HADS-D. Therefore. 11 on the GDS-15. However when pooling studies resulted in I2 90%. Additional subgroup analyses were conducted for older adults. 1994). and diagnostic odds ratios. PsycINFO. the most consistently reported and recommended cut-off points for each of the scales were extracted (see Table 7). There are three main instruments that 105 . 2007) commands in order to obtain pooled estimates of sensitivity. 17 on the CES-D. a higher threshold for acceptable heterogeneity in such meta-analyses is required. six on the ‘Whooley questions’. positive likelihood. two on the GHQ-28. likelihood ratios and diagnostic odds ratio (for further details. and any one.Case identification and service delivery Table 6: Databases searched and inclusion/exclusion criteria for the effectiveness of case identification instruments Electronic databases Date searched Study design Patient population Instruments MEDLINE. five on HADS-total and seven on one-item measures (see Appendix 20 for further details). 12 on the GHQ-12. To maximise the available data.. see Chapter 3). CES-D. GHQ. specificity. EMBASE. GDS. meta-analyses were not conducted. community. Zung Self Rated Depression Scale. 20 on the GDS. In addition. negative likelihood Outcomes were on the PHQ-9. specificity. diagnostic odds ratio. Cochrane Collaboration. 19 on the BDI. 2007. 251 studies were excluded from the analysis. specificity.or two-item measures of depression Sensitivity. positive likelihood ratios. AUC. HADS. Cochrane Library Database inception to February 2009 Cross-sectional studies People in primary care. Evaluating identification tools for depression A bivariate diagnostic accuracy meta-analysis was conducted using Stata 10 with the Module for Meta-analytical Integration of Diagnostic Test Accuracy Studies (MIDAS) (Dwamena. and general hospital settings BDI. Heterogeneity is usually much greater in meta-analyses of diagnostic accuracy studies compared with RCTs (Gilbody et al.. negative likelihood ratios. Table 8 summarises the results of the meta-analysis in terms of pooled sensitivity.

Gilbody et al. 0. Although the PHQ-2 and the Whooley questions use the same two items. 2002.. 95% CI. the Whooley version dichotomises the questions (yes/no) and has a cut-off of 1 compared with 3 for the PHQ-2. 0. For the PHQ-9 in consultation samples (people in primary care or general medical settings) there was relatively high heterogeneity (although of a similar level to most other scales) (I2 74. 95% CI. This scale was found to have high sensitivity (0.95.76).88). The PHQ-2 could not be meta-analysed as there was very high heterogeneity.55. 95% CI.. PHQ-2 (Kroenke et al.83. 1997). 0.86) and specificity (0. 11–14 moderate. 1999). 0. The PHQ-9 has nine items and has a cut-off of 10.82.91.. A single study by Arroll and colleagues 106 . 15 16 10 5 ? 50 mild. 0. 2007) Scale BDI 21 items 13 items Primary care version PHQ 9 items 2 items 2 items (Whooley version) GHQ 28 items 12 items HADS-D CES-D GDS 30 items 15 items 5 items Zung Cut off points 13 4 4 10 3 1 5 3 8–10 mild.77. 2003) and the ‘Whooley questions’ (Whooley et al. 60 moderate. 0. 0.04%).97) but lower specificity (0.. the difference is that while the PHQ-2 follows the scoring format of the PHQ-9 (Likert scales).. 0. 70 severe severe have been developed from this scale.66. 95% CI. The PHQ-9 was found to have good sensitivity (0.Case identification and service delivery Table 7: Cut off points used (if available) for each of the identification tools (adapted from Pignone et al. the PHQ-9 (Spitzer et al.76. The Whooley questions analysis included studies both on consultation and chronic physically ill samples as there were too few studies to break down by population.

87) 0.66 (0.97) 0.27) 18. 0.25 (0.67.02 (1. 0. 0. 6.87 (0.79) 0.17.29) 15.85. 0.92) BDI-non somatic items Consultation sample: 5 studies 0.12. 0. 0. 0.09.85 (0. 33.83. 3. 0.89) 0.01.82. 0.83 (0.92 (3.79 (0.65. 17.77.76) 2. 0.86) 0.88) 4. 4.90 (0.89) GDS-15 Consultation sample: 11 studies 0.22) 0.35.81 (0. 24.81.17. 0.15.04. 0.22) 0. 0.78. 0.76.67 (5.79.29 (15.91) 0.88) 1-item Consultation sample: 6 studies 0.73) 2.83.60) 0. 4.87.25 (14.85 (0.84 (0.94 (0.90) 0.19 (2.46) 107 was not possible to conduct separate subgroup analyses for consultation and chronic physical illness samples due to lack of studies for the Zung and Whooley questions.65 (0. 0.24) 29. 0.24) 0.05) 0.40 (2.74 (0.83 (0.94) 0. 56. 0. 0.04) 0. 0.88 (0.57.22 (0. 0.52) Whooley*: All populations: 7 studies 0.36) 9.84 (0. 0.02.91. 6.80. 31. 0. 3.89.83) 3.21 (0. 0.82 (2. 4.91) Likelihood ratio Likelihood ratio Diagnostic odds ratio AUC Population and instrument Sensitivity PHQ-9 Consultation samples: 11 studies 0.29.18 (0.55.29) 21.80) 3.87. 0.35. 0.61.13) 0. 0.08 (0. 0. 0.82 (0.Table 8: Evidence summary of depression identification instruments in primary care.55.79. 0.95 (0.87.96) BDI Consultation samples: 4 studies 0.17.38 (1.69) 11.91) 5. *It .90) 0.83 (0.78.73 (0.73.38 (11.89) 0. 0.80.25 (0.72) 0.38.75 (0.82 (2.85. people with a chronic physical health problem.74.92. 47.86) 0.70.12. 9.33) 15.24 (0.95 (8.41.96) 0.14 (2.02 (9.86 (0. 0.20) Case identification and service delivery 0.70 (3.87 (0. 88.86) CES-D Consultation sample: 8 studies 0.32) 0.24) 0.18 (0. 38.69.10.98 (10. 0.83 (0.88) 3.90) Older adults: 5 studies 0.15) 36.05.81) 3. and older populations Specificity 0.84.82 (0.

1961) and subsequently updated the original 21-item version (Beck et al.85. It was not possible to a conduct meta-analysis on the effects of any of the PHQ scales or the Whooley questions on older adults because of a lack of data (one study each on the PHQ-9. This was also consistent with the diagnostic odds ratio (29. For the 21-item BDI there was high heterogeneity for consultation samples (I2 88. 1979. 95% CI. 1991) was developed as a general measure of psychiatric distress and measures a variety of constructs such as depression and anxiety. consistency between studies to assess these scales (BDI: non-somatic) in consultation (I2 75. In addition.73.91). commonly. The BDI appeared to perform relatively well in terms of sensitivity (0. 95% CI. There was sufficient. while there were more studies on the GHQ-12 there was very high heterogeneity (I2 90%) for studies on consultation populations. There was high sensitivity (0... 0. 0. This scale has been used widely as a depression outcome measure and is also used to provide data on the severity of depression. 1997). but as there was only a single study showing the effect of this approach the GDG decided not to adopt it. therefore these studies were also not meta-analysed.90) and specificity (0.79). Beck et al. Beck Depression Inventory Beck originally developed the BDI in the 1960s (Beck et al. 56. General Health Questionnaire The GHQ (Goldberg & Williams... However. a meta-analysis specifically for older adults was not possible due to there being only two studies. 13 is used a cut-off in identification studies. 0. The main versions used for identification purposes are the GHQ-28 (cut-off of 5) and GHQ-12 (cut-off of 3).61%).57. this is based on only four studies so it is difficult to draw firm conclusions. 0. asking the patient if they wanted help with their depression. 1974. 2000) and BDI short-form (Beck et al. although relatively low.. 1996). 95% CI.83).79.71%) populations. 0. the BDI-fast screen has been specifically developed for use in primary care (Beck et al. 0. 95% CI. A meta-analysis was not possible for older adults as there were only two studies. 108 . Subgroup analyses on older adults were also not possible as there were only two studies for this population.94) but lower specificity (0.61.Case identification and service delivery (2005) added a further question to the two in the PHQ-2.70. In addition.83. 0. 0. PHQ-2 and Whooley questions). Moreover.. therefore meta-analysis was not conducted. 95% CI. 1996) were combined to assess the impact of removing somatic items as data from both scales were relatively sparse. 15. Furthermore. There were only two trials of the GHQ-28. the cognitive–affective subscale of the BDI has often been used to identify depression.82.29. This increased specificity and the GDG considered the findings of the study and the adoption of the third question. Beck Depression Inventory – non-somatic items Data from BDI fast-screen (Beck et al.103.

80. 0.89) but very low specificity (0.85.96%). 95% CI. 1976). This measure is also relatively commonly used as an outcome measure. One-item measures Five studies were found to assess a one-item measure in consultation samples. however meta-analysis could not be conducted due to very high heterogeneity (I2 90%) for all subgroups including consultation populations and older adults.65.79. 95% CI. there was very high heterogeneity (I2 90%) for the GDS. 0.55. 0.63%) sample. However. 0. For consultation samples sensitivity was high (0. The diagnostic odds ratio indicated a lack of accuracy 109 . 95% CI. The depression subscale has seven items and the cut-off is 8 to 10 points. There are various short forms of the CES-D including an eight-. Center for Epidemiological Studies Depression Scale The CES-D (Radloff. 95% CI.73). It is sometimes used as an outcome measure as well.65.74.20). 33.78. 0. 10. 1977) has 20 items and the cut-off is 16. 1983) is a measure of depression and anxiety developed for people with physical health problems.69. 0.78. 0. 0. 0. No subgroup analyses for older people were conducted as all participants were over 65 years of age.87.67.87) and higher specificity (0. In the consultation population there was higher sensitivity (0.74.and 11-item scale. 95% CI. it was possible to analyse studies on the GDS-15. 1965) and has been revised (Guy.91) but specificity was relatively low (0. 95% CI. There was a relatively good sensitivity (0.81. For the older adult population. The original 30-item scale (cut-off of 10 points) was developed by Yesavage and colleagues (1983) and more recently a 15-item (cut-off of 5 points) version has been validated. A total of 21 studies were included in the review. there was relatively low sensitivity (0. 0. 0.75. 0. For older adults.84. ten. 95% CI. 95% CI. 0. There was high heterogeneity in the consultation (I2 84. 0. Heterogeneity was relatively acceptable (I 2 70.84. Geriatric Depression Scale The GDS was developed to assess depression in older people. 95% CI. The diagnostic odds ratio was just below 20 (18.Case identification and service delivery Hospital Anxiety and Depression Scale The HADS (Zigmond & Snaith. Zung Self-Rating Depression Scale The self-rating depression scale was developed by Zung (Zung.87). 0. 0. therefore no meta-analyses could be conducted. There were insufficient studies to conduct a meta-analysis.98.81). Haringsma and colleagues (2004) was removed from the analysis resulting in acceptable heterogeneity (I2 61.09%).89) but specificity was lower (0. Despite the large number of studies (18 studies). This has 20 items where a cut-off of 50 is typically used.80).

85 0.23 1.12 1. The GDS and GDS-15 were almost always used for older adults.73 0. 95% CI.02 0. Zung.09 0. Comparing validity coefficients for case identification tools in older adults The impact of old age and residing in a nursing home on the validity coefficients of the case identification tools reviewed above were assessed through meta-regression (see Table 9).34 0. and one-item measures were not included in the analysis.65 0.32 Joint I2 0 Joint I2 58.58 2.79 0.67.13 2.44 0.18 0. therefore the validity of these measures in older adults is already accounted for in the previous analysis.28 .14 0.35.74 1. 5. Because of a lack of data the PHQ-2.80 0.33 0. Whooley.58 0. Table 9: Meta-regressions assessing the impact of differences within populations of studies Population and instrument PHQ-9 Comparing over 65s with under 65s BDI Comparing over 65s with under 65s BDI-non somatic items Comparing over 65s with under 65s CES-D Comparing over 65s with under 65s GDS Comparing nursing home with non-nursing home GDS-15 Comparing nursing home with non-nursing home GHQ-12 Comparing over 65s with under 65s 110 Beta-coefficient Sensitivity Specificity Sensitivity Specificity Sensitivity Specificity Sensitivity Specificity Sensitivity Specificity Sensitivity Specificity Sensitivity Specificity 1.80 0.34 0.Case identification and service delivery (9. It was not possible to conduct a subgroup analysis of older adults as there were only two studies.43 1.45 I2(%) Joint I2 0 p-value 0.46).91 0. 17. further analyses were conducted to assess the validity of these measures in nursing home populations.09 0.83 0.36 0.23 1.84 1. However.54 1.64 Joint I2 43.17 0.65 0.30 Joint I2 0 Joint I2 0 Joint I2 11.14 0.61 1.65 0.

For example... For all other scales there were no statistically significant differences. whereas a Healthcare Commission survey highlighted how both Asian and black/black British people were less likely to be offered ‘talking therapies’ (Department of Health. are more likely to present to their GP with physical manifestations. including both psychological and pharmacological interventions.02) were associated with improved specificity in older adults compared with people under 65 years. Bangladeshi and Pakistani people. There is an increasing evidence base to suggest that the reduced identification of depression in different ethnic and cultural groups may be one barrier to receiving appropriate treatment. 2008b). Weich et al.09) in older adults compared with people under 65 years. both Wilson and MacCarthy (1994) and Williams and Hunt (1997) have indicated that despite this increased GP contact. socioeconomic status and perceptions of disadvantage (Bhugra & Cochrane. 2004).Case identification and service delivery Older adults There was some evidence that the BDI versions with no somatic items (p 0. research has suggested that across mental disorders. Such findings cannot wholly be explained by differences in factors such as urbanicity. Shaw and colleagues (1999) indicated that women from black and minority ethnic groups had an increased incidence of common mental disorders including both depression and anxiety. a number of findings have indicated both ethnic and cultural variations in the subjective experience and initial presentation of the illness. including common mental disorders such as depression (Bhui et al. culture is known to exert an influence on the presentation and subjective experience of illness. People in nursing homes Only the GDS and GDS-15 provided sufficient data on people in nursing homes.2. and do so more frequently than their white counterparts. Department of Health. GPs are less likely to detect depression and more likely to diagnose ‘Asians’ with a physical disorder. For example. Commander and colleagues (1997) are among researchers who suggest that ‘Asians’. There appeared to be limited differences in validity when assessing people either in nursing homes or in the community for both scales. What a person perceives as an illness and whom they seek for treatment are all affected by their culture and ethnicity. and even when a psychological problem is present. 2001). 2008b). there was often a lack of power in most studies because only a small number of studies on older adults were found for most scales. Furthermore. 2003.. With regard to depression. There was a trend towards reduction in sensitivity for the CES-D (p 0. However. 2001. including Indian. However.9 Case identification in black and minority ethnic populations Introduction Culture and ethnicity are known to influence both the prevalence and incidence of mental illnesses. 5. For example. 111 . particular ethnic groups are often under-represented in primary care services (Bhui et al.

most of which are still early in their development. aim to incorporate specific cultural idioms and descriptions commonly reported by people from a particular ethnic or cultural group. Information about the databases searched and the inclusion/exclusion criteria used are presented in Table 10. Such tools. 2010). the decision was made to only include UK studies.Case identification and service delivery Despite an increased awareness that different cultural and ethnic factors may influence the presentation of depression. 12Papers assessing the validity of established scales in UK black and minority ethnic populations were required to have a ‘gold standard’ diagnosis defined as DSM–IV or ICD–10 diagnosis of depression. in a UK black and minority ethnic population. namely people of Pakistani family origin. As discussed above. Owing to the above evidence indicating ethnic and cultural variations in the presentation and subjective experience of illness. All four papers were conducted within the community or primary care. the presentation and subjective experience of depression is known to be influenced by cultural and ethnic factors. Studies considered A total of four studies met the eligibility criteria of the review. the majority of case identification tools used in routine clinical practice were originally created and validated in white populations (Husain et al. This was undertaken as a joint review for this guideline and the guideline for depression in adults with a chronic physical health problem (NCCMH. the Personal Health Questionnaire. and two studies compared the Caribbean Culture-Specific Screen for emotional disorders (CCSS) with the GDS. Databases searched and inclusion/exclusion criteria The review team conducted a new systematic search for cross-sectional studies aiming to assess tools for identifying depression. One included study compared the Amritsar Depression Inventory (ADI) with the GHQ-12. 112 . The review also assessed the validity of established depression case identification tools for different black and minority ethnic populations within the UK12. Only one study assessed the validity of an established scale. it was felt that findings from non-UK ethnic minority populations would not be generalisable because of the ethnic and cultural differences among the populations studied.. Definition and aim of topic of review The review considered any ethnic-specific case identification instruments aimed at detecting depression in black and minority ethnic populations. therefore. 2007). one proposed method to improve the identification of depression in black and minority ethnic participants is to assess the validity of ethnic-specific screening tools. Details of the search strings used are in Appendix 8. Although the GDG was aware of papers from outside the UK (most notably from the US). This included new identification tools designed for different cultural and ethnic groups. and also existing scales modified and tailored towards the specific needs of particular black and minority ethnic groups.

Bhui and colleagues (2000) screened both Punjabi and white English attendees of five primary care practices in South London. assessed in a UK black and minority ethnic population Sensitivity. CES-D. PHQ. Any of the above validated identification tools. specificity. The most common reason for exclusion was that the paper was a non-UK based study/population or that the paper presented no usable evaluation of a screening tool. and any one. Evaluating identification tools for depression in black and minority ethnic populations Because of both the paucity of data on ethnic specific scales in the UK and differences in the populations and instruments investigated. Any cultural or ethnically adapted version of the following validated case identification instruments: BDI. Throughout the study.or two-item measures of depression 3. a cultural screen assessing self-affirmed cultural origin was 113 . Cochrane Library Database inception to February 2009 Cross-sectional studies People in primary care. Using the ADI and the GHQ-12. 1974). diagnostic odds ratio. it was not possible to conduct a metaanalysis of the included studies. GDS. positive likelihood. GHQ. Amritsar Depression Inventory The ADI is a culturally specific instrument developed in the Punjab in India and is aimed at detecting depression in the Punjabi population of the Indian subcontinent (Singh et al. and general hospital settings from black and minority ethnic groups 1. community. Instead the findings from the included studies are summarised in a narrative review below.. ten studies were excluded from the analysis.Case identification and service delivery Table 10: Databases searched and inclusion/exclusion criteria for clinical effectiveness of psychological interventions Electronic databases Date searched Study design Patient population Instruments MEDLINE. AUC. negative likelihood Outcomes In addition. The 30-item dichotomous (yes/no) questionnaire was developed on the basis of 50 statements commonly used by Punjabi people with depression. HADS. The screen development process also utilised frequently used ‘illness statements’ and common descriptions of signs and symptoms of depression prevalent in the psychiatric literature. Zung Self Rated Depression Scale. EMBASE. PsycINFO. Any ethnic-specific depression case identification instrument 2.

the ADI performed worse in detecting depression in the Punjabi participants. During this second stage interview. although this finding was not sustained for the GHQ-12 in which the same cutoff was optimal for both groups. One additional finding of interest was that the optimal cut-off for the ADI was higher for the Punjabi participants compared with their white English counterparts. depending on the preference of the participant. the screening tools were administered in English. Results of the validity coefficients and ROC curve analysis using the standard CIS-R thresholds for depression indicated that while the GHQ-12 performed well across both groups. Participants were firstly administered the CCSS. culture had an impact on the validity coefficient of the ADI. Two papers assessed the validity of the CCSS screen in older African–Caribbean participants living in two different locations in the UK. A two-phase screening protocol was applied in which all ‘probable cases’. Both papers compared the validity of the CCSS to the GDS and utilised the Geriatric Mental State-Automated Geriatric Examination for Computer Assisted Taxonomy (GMS-AGECAT) as a gold standard for case identification. In particular. GDS-15 and the Mini-Mental State Examination (MMSE). namely South London and Manchester. which was informed through a process of consultation with African–Caribbean religious healers/ministers. 114 . were administered to the selected participants. Barbados. Trinidad and Grenada. The sample in Abas and colleagues (1998) consisted of consecutive AfricanCaribbean primary care users aged over 60. Results indicated that the ADI was no better than chance in identifying cases of depression. Responders were categorised as high scorers if they scored 4 on either measure. Analysis of the individual items of both the GHQ-12 and the ADI failed to indicate any specific items that were strongly predictive of depression caseness in either cultural group. the GMS-AGECAT and a culturally-specific diagnostic interview. and low scorers if they attained less than 4 on both screens. particularly for Punjabis who had been resident in the UK for more than 30 years. Caribbean Culture-Specific Screen for emotional distress The CCSS (Abas. Punjabi or a combination of the two. for example. although performing in line with the GHQ-12 for the white English participants. and one third of ‘probable non-cases’ proceeded to a second interview in which the Clinical Interview Schedule-Revised (CIS-R) was administered by a bilingual psychiatrist. To overcome any additional barriers because of language. 1996) is a 13-item dichotomous (yes/no) culture-specific screen which was developed through a process of generating locally-derived classifications of mental disorders in Caribbean people and gathering commonly used terms for emotional distress.Case identification and service delivery applied to detect both Punjabi and white English participants. and included both clinic attendees and those receiving home visits from primary care teams. The majority of participants interviewed in the piloting stages of the screen were from Jamaica with a number of participants identifying themselves as from other Caribbean countries including Guyana. those scoring 2 on the GHQ or 5 on the ADI. A random sample of 80% of the high scorers and 20% of the low scorers was selected to attend a further interview.

and no suggestions for changes were made. the Personal Health Questionnaire contains no ‘difficult culture specific idioms’. particularly when delivered in a culturally sensitive way. Although slight variation existed between the two analyses. This conclusion was supported by Abas and colleagues (1998) who found that a proportion of participants were more likely to discuss and disclose information during the culturally sensitive diagnostic interview. During this stage. In the Abas and colleagues’ (1998) paper it was demonstrated that at a certain cut-off the GDS appeared to perform better than the CCSS. rather than the content per se. thus making translations into other languages possible. for example. the cultural competence of staff and delivering the screen in a culturally sensitive way. All included participants were interviewed by one of two interviewers of a similar cultural background. the culturally-specific CCSS did not outperform the GDS. Jamaica. In the present study. In both papers. Because it was noted that considerable variation may exist among people of Caribbean origin from different islands. although the authors noted that the small sample size prevented any meaningful test of statistical significance. namely the GDS-15. both papers have suggested that routine clinical screens may be appropriate for black and minority ethnic participants. Rait and colleagues (1999) argue that the success of case identification measures may be more dependent on the way in which the screen is delivered. Personal Health Questionnaire Husain and colleagues (2007) assessed the validity of the Personal Health Questionnaire in Pakistani people who were resident in the UK. Consequently. Registers for general practices with a high-proportion of African–Caribbeans were used to identify members of the community. In stage one. The ROC curve analyses for the papers indicated that both the GDS and the CCSS performed well in the populations. unlike many screening instruments. the results of Rait and colleagues’ (1999) paper were presented for the sample as a whole and for a subgroup of Jamaican people who constituted the majority of participants. Trinidad and so on. with group discussion utilised to reach a single consensus. for example. with the same optimal cut-off occurring in both analyses. the main language of immigrants from Pakistan. 115 . the results were similar. CCSS and the Brief Assessment Schedule Depression Cards (BASDEC).Case identification and service delivery Rait and colleagues (1999) included a community sample of African–Caribbean people aged 60 years and over. with those who consented to take part in the study subsequently interviewed in their homes. One important feature of the Rait and colleagues’ (1999) study was that the authors sought advice from a panel of community resident African–Caribbeans regarding the acceptability of the GDS. The authors noted that. used as a diagnostic ‘gold standard’ for the detection of depression. The content of the screens was deemed acceptable. with a high level of sensitivity and specificity when using the GMS-AGECAT as a gold standard for diagnosis. letters were sent to potential participants. three depression screens were applied. when compared with the standard GMSAGECAT. the Personal Health Questionnaire was translated and back-translated into Urdu. The second stage of the study involved the home administration of the GMS-AGECAT.

who may or may not be wholly representative of ethnic minorities.2. 2000). 5. Further research into this area is therefore required to answer such questions. The high sensitivity and specificity at the recommended cut-off suggested that the Personal Health Questionnaire is able to detect depression in people of Pakistani family origin when administered in either English or Urdu. which is an important limitation of the reviews. All eligible participants first completed the Personal Health Questionnaire in either English or Urdu. A further caveat to consider is that three of the four studies that were included assessed consecutive primary care attendees. Furthermore. particularly those who experience barriers to accessing and engaging with primary care services.. Limitations with the evidence base It must be noted that a number of potential limitations exist in relation to the above studies. Only one paper used a culturally-sensitive diagnostic tool as a measure of caseness (Abas et al. Eligible participants were identified through either their name and/or language or via direct questioning. the authors noted that participants in this study and in a study conducted in Pakistan (Husain et al.6 and a negative predictive value of 80. However. the findings of one paper in which a community sample was recruited were consistent with the results of the primary care studies. a semi-structured interview resulting in an ICD diagnosis. but are instead based on ethnocentric ideas of mental illness (Bhui et al.4% and a specificity of 89.6. When compared with the Whooley 116 .10 Clinical summary for both reviews There was very high heterogeneity found for almost all identification tools. a two stage process was employed. suggesting the findings may be robust for each particular ethnic group under investigation. The remaining three papers compared the screens with long-standing measures predominantly based on the DSM and ICD–10 classification systems. number of items and scoring systems. with a research psychiatrist administering the screen in the case of illiteracy. all participants were interviewed in either their home or within the primary care practice.. One caveat is the lack of an established gold standard for the diagnosis of depression in people from black and minority ethnic groups. 1998). Results of the ROC curve analysis indicated that the recommended cut off score of 7 produced a sensitivity of 70. Consequently. It is argued that these measures may not be culturally specific and sensitive to cultural differences. with a positive predictive value of 82.Case identification and service delivery Consecutive primary care attendees of Pakistani family origin aged 16 to 64 years were included in the sample. any culturally sensitive measure may not be expected to have a high sensitivity and specificity for caseness when compared with these diagnostic measures.3%. In the second stage of the study. 2000) did not experience any difficulties in understanding and answering the screening questions. Scales varied a great deal in terms of targeted populations. in either Urdu or English dependent on preference. depending on patient preference. A psychiatrist administered the Psychiatric Assessment Schedule. As with the other screening studies..

which included scales specifically targeted at this population (for example. namely Pakistani people resident in the UK. the GDS and GDS-15) as well as all other measures reviewed. failed to detect depression in different ethnic and cultural groups. The first stage of case identification would require using a highly sensitive instrument that could be used in routine clinical practice with limited training and implementation difficulties. Section 3. other scales such as the PHQ-9 and GDS-15 had better specificity but not as much sensitivity (although they still met the criteria for high sensitivity). the number of studies was too small to conduct metaanalyses for any of these measures. There were studies that targeted older adults for all of the other scales reviewed. the GDG concluded that in the first stage of case identification the Whooley questions remained an appropriate tool for depression.11 Health economic evidence and considerations No evidence on the cost effectiveness of case identification tools for depression in primary care and community settings was identified by the systematic search of the economic literature. 5. given the lack of specificity found with the Whooley questions it was the view of the GDG that people with a positive response would benefit from a more detailed clinical assessment. Details on the methods used for the systematic search of the economic literature are described in Chapter 3. however.2. There were also planned subgroup analyses conducted for older adults. which may include a more detailed instrument possessing better overall psychometric properties. this was validated in a particular black and minority ethnic group. There was a paucity of data concerning ethnic-specific identification tools. given that this measure is already in current use in primary care. as well as the possible benefit of a two-stage process of identification and diagnosis. In all studies. 5.6. However.Case identification and service delivery questions. but the available data suggested a slightly (although not statistically significant) reduced sensitivity compared with consultation populations as a whole.12 From evidence to recommendations The GDG noted the different nature of the scales contained in the review and their psychometric properties. There were fewer studies on the CES-D. validated and well researched measures such as the GHQ-12 outperformed the ethnicspecific scales in terms of both sensitivity and specificity.1. which may be in their developmental infancy.2. with limited data suggesting that the scales. However. the large number of studies on the 30-item GDS could not be meta-analysed as there was very high heterogeneity. in the case of the Personal Health Questionnaire. Furthermore. The data on case-finding instruments in black and minority ethnic groups did not identify any specific measures that in the opinion of the GDG improved upon the results obtained 117 . The data supported the use of the Whooley questions and. The GDS-15 appeared to be relatively effective in consultation populations.

for example people with sensory impairments or a learning disability.3 5. The person places a mark on the scale answering: ‘How distressed have you been during the past week on a scale of 0 to 10?’ Scores of 4 or more indicate a significant level of distress that should be investigated further (Roth et al.Case identification and service delivery with the Whooley questions. If a person answers ‘yes’ to either of the depression identification questions (see 5.13 5. a practitioner who is competent to perform a mental health assessment should review the person’s mental state and associated functional.13. Other recommendations from the previous guideline remain essentially the same. depressed or hopeless? ● During the last month. for symptoms.2.13. have you often been bothered by having little interest or pleasure in doing things? If a person answers ‘yes’ to either of the depression identification questions (see 5. and therefore no specific black and minority ethnic recommendations on case finding tools are made. the need for cultural competence of staff in assessments was noted in the review of case-finding instruments in black and minority ethnic groups.13. they should refer the person to an appropriate professional. self-harming or over-active behaviours or comorbid physical health problems such as epilepsy. have you often been bothered by feeling down.13. diabetes or heart disease (Prasher. as recommended in the previous guideline. 5.13. will be under-diagnosed. specifically: ● During the last month. inform the GP of the referral.2. For people with significant language or communication difficulties. established aggressive. If this professional is not the person’s GP.1 Recommendations Be alert to possible depression (particularly in people with a past history of depression or a chronic physical health problem with associated functional impairment) and consider asking people who may have depression two questions.4 5.2. 118 . When assessing a person with suspected depression. the need to move beyond simple symptom counts was noted.2. consider using a validated measure (for example.2. and this is reflected in the recommendations.2.2. 1999. so the recommendation from the previous guideline has been amended.1). In addition.2 5. 1999) in this population. a learning disability.. 2007).13. 1998). interpersonal and social difficulties. This guideline update also makes recommendations for people with depression and learning disabilities or acquired cognitive impairments because it is likely that depression. However.13. in performing a more comprehensive mental health assessment. functions and/or disability) to inform and evaluate treatment.2.5 13The Distress Thermometer is a single-item question screen that will identify distress coming from any source. which is ‘relatively common’ (Prasher.1) but the practitioner is not competent to perform a mental health assessment. particularly if they have autism. consider using the Distress Thermometer13 and/or asking a family member or carer 5. Mind.

In addition to assessing symptoms and associated functional impairment. treatments ● the quality of interpersonal relationships ● living conditions and social isolation. provide the same interventions as for other people with depression ● if necessary. uk/CG38). be aware of any learning disabilities or acquired cognitive impairments and.2.10 When depression is accompanied by symptoms of anxiety.8 When assessing a person with suspected depression.2. 14Refer if necessary to ‘Bipolar disorder’ (NICE clinical guideline 38.7 Learning disabilities 5.nice.2. When the person has an anxiety disorder and comorbid depression or depressive symptoms. conduct a comprehensive assessment that does not rely simply on a symptom count.13. consult the NICE guideline for the relevant anxiety disorder and consider treating the anxiety disorder first (since effective treatment of the anxiety disorder will often improve the depression or the depressive symptoms)15.13.13.Case identification and service delivery about the person’s symptoms to identify possible depression. Any wording changes have been made for clarification only. and response to. If a significant level of distress is identified. consider how the following factors may have affected the development. adjust the method of delivery or duration of the intervention to take account of the disability or impairment. investigate further. 119 .6 5.org. available at www.13.13. course and severity of a person’s depression: ● any history of depression and comorbid mental health or physical disorders ● any past history of mood elevation (to determine if the depression may be part of bipolar disorder14) ● any past experience of. 5. consider consulting with a relevant specialist when developing treatment plans and strategies. 5.2. Depression with anxiety 5. 15The evidence for this recommendation has not been updated since the previous guideline. Take into account both the degree of functional impairment and/or disability associated with the possible depression and the duration of the episode.9 When providing interventions for people with a learning disability or acquired cognitive impairment who have a diagnosis of depression: ● where possible. When assessing a person who may have depression. if necessary.2. the first priority should usually be to treat the depression.

This is particularly important during high-risk periods.2. 17Ibid. in discussion with stakeholders and with the GDG. In the process of the development of this guideline.13. and suicidal ideation. 18Ibid. 5. 5. the GDG preferred the term ‘active monitoring’ and revised the original recommendation accordingly.11 Always ask people with depression directly about suicidal ideation and intent.13 Advise people with depression of the potential for increased agitation. negativity and hopelessness.Case identification and service delivery Risk assessment and monitoring 5. anxiety and suicidal ideation in the initial stages of treatment. considerable concern was expressed about the term itself and the fact that it suggested a passive process rather than the more active process of assessment. Active monitoring In the previous guideline.2. actively seek out these symptoms and: ● ensure that the person knows how to seek help promptly ● review the person’s treatment if they develop marked and/or prolonged agitation.13. refer them urgently to specialist mental health services17. and to contact their practitioner if concerned. In light of this. may recover with no formal intervention.13. such as starting or changing treatment and at times of increased personal stress18.2. 5.16 For people who. such as more frequent direct or telephone contacts ● consider referral to specialist mental health services19.13. If there is a risk of self-harm or suicide: ● assess whether the person has adequate social support and is aware of sources of help ● arrange help appropriate to the level of risk ● advise the person to seek further help if the situation deteriorates16. 5.15 If a person with depression is assessed to be at risk of suicide: ● take into account toxicity in overdose if an antidepressant is prescribed or the person is taking other medication. 120 . advice and support that characterises effective interventions for people with mild depression that may spontaneously remit.12 If a person with depression presents considerable immediate risk to themselves or others. 19Ibid.2.13. in the judgement of the practitioner.14 Advise a person with depression and their family or carer to be vigilant for mood changes.13. limit the amount of drug(s) available ● consider increasing the level of support. a recommendation was made for watchful waiting.2. or people with mild depression who do not want an 16Ibid.2. if necessary. 5.

2002). Referral to specialist care is only expected to be required in a small proportion of cases. This may involve ‘graduated access’ to services. 2005a) as a component (Katon et al.. However. in order to support them in caring for specific patients who are currently undergoing care. in this model specialists enter into an ongoing educational relationship with the primary care team. Bower & Sibbald. rather than providing training interventions that teach skills in dealing with patients with depression in general. The consultation-liaison model This model (for example. 1989. the over-arching term ‘enhanced care’ is used to refer to them all. In this guideline update. normally within 2 weeks ● make contact if the person does not attend follow-up appointments. Graduated access One way of changing access is to modify service provision at the point at which people want to access services (Rogers et al. or people with subthreshold depressive symptoms who request an intervention: ● discuss the presenting problem(s) and any concerns that the person may have about them ● provide information about the nature and course of depression ● arrange a further assessment. This includes a number of interventions or models that often have some degree of overlap or where individual interventions are contained within large models. 2000). Darling & Tyler. Creed & Marks.1 As indicated above.3.Case identification and service delivery intervention. The attached professional model In this model (for example. 2006) may include stepped care (Bower & Gilbody.. 1990.. in that it seeks to improve the skills of primary care professionals and improve quality of care through improvements in their skills. Unutzer et al. including the use of ‘direct health services’.3 SERVICE DELIVERY SYSTEMS IN THE TREATMENT AND MANAGEMENT OF DEPRESSION Introduction 5. which people can access without having face-to-face contact with professionals and which maximise the use of new technologies such as the internet. 5. For example. 1999... 1999). Some of the more prominent models are listed below. A common implementation of this model involves a psychiatrist visiting practices regularly and discussing patients with primary care professionals. 1997) is a variant of the training and education model (which is outside of the scope of the guideline). Gask et al. there has been a considerable number of service-focused developments since the publication of the previous guideline. a mental health professional has direct responsibility for the care of a person (usually in primary care) focusing on the 121 . collaborative care interventions (Gilbody et al.

Stepped care Stepped care (for example. 2006). 1997. it may also involve the co-ordination of follow-up. Van Straten et al. Bower & Gilbody.to high-intensity interventions. although in other systems patients are stepped up to a higher intensity intervention on immediate contact with the service. in the context of patient preference and readiness ● the creation of a range of self-management training and support services in which patients have access to services that teach the necessary skill to carry out treatment plans. Contact is usually limited to treatment and involves little or no follow-up beyond that determined by the specific intervention offered (for example. 2001) emerged from the chronic disease model and has four essential elements. for example if they are acutely suicidal (this later model is the one adopted in this guideline update and in the previous guideline). Collaborative care The collaborative care model (for example. be it pharmacological or psychological. identify possible complications and check and reinforce progress in implementing the care plan. if a person does not benefit from an initial intervention they are ‘stepped up’ to a more complex intervention). and which has a self-correcting mechanism built in (that is. and the consequent treatment is determined by the assessing professional in consultation with the patient. Typically. goals and plans are jointly developed by the patient and professional to achieve a reasonable set of objectives.Case identification and service delivery primary treatment of the problem/disorder. booster sessions in CBT). 122 . 2005a) is a system for delivering and monitoring treatment with the explicit aim of providing the most effective yet least burdensome treatment to the patient first. moving from low.. Case management This describes a system where an individual healthcare professional takes responsibility for the co-ordination of the care of an individual patient (for example. Stratified (or matched care) This is a hierarchical model of care (for example. where at the patient’s point of first contact with services they are matched to the level of need. 2006). stepped care starts by providing low-intensity interventions. In some stepped-care systems. low-intensity care is received by all individuals. Wagner. which are: ● the collaborative definition of problems. in which patient-defined problems are identified alongside medical problems diagnosed by healthcare professionals ● a focus on specific problems where targets. Katon et al. but is not necessarily directly involved in the provision of any intervention. The co-ordination of care remains with the GP/primary care team. collaborative care also typically includes a consultation liaison role with a specialist mental health professional and generic primary care staff. Gensichen et al. In mental health services... guided behaviour change and promote emotional support ● the provision of active and sustained follow-up in which patients are contacted at specific intervals to monitor health status.

the development of stepped care (Davison. One consistent factor that links these developments is the limited evidence for most if not all of these interventions. 2000). In this guideline it is assumed that collaborative care.Case identification and service delivery It may also include elements of many of the other interventions described above. developments in the treatment of depression. Perhaps the model most widely adopted has been the stepped-care model within the IAPT programme (Department of Health. There will be joint determination with the service user regarding the care plan along with long-term coordination and follow-up. 2006) there has not been a consistent adoption of any particular model of stepped care. focused on the treatment and care of depression. A significant factor in driving these developments has been the recognition that for many people depression is a chronic and disabling disorder. 5. 2007). and influences from physical healthcare (for example. which has grown considerably in the past 10 years and has led some (for example. care must be taken when considering its adoption in different healthcare systems (Campbell et al. but there have also been changes in mental healthcare policies that have influenced implementation.3. there has been a growing interest in the development of systems of care for managing depression. The implementation in the NHS of the various developments described in the introduction has been variable. 2002]). The most notable exception is the evidence base for collaborative care. All sectors of care should be involved in order to ensure a comprehensive and integrated approach to mental and physical healthcare. is provided as part of a well-developed stepped care programme. Typically the programme of care is coordinated by a dedicated case manager supported by a multi-professional team. chronic disease management [Wagner & Groves. However. psychological and pharmacological interventions.2 Current practice and aims of the review Over the past 20 years. such as managed care and Health Maintenance Organisations (Katon et al. Simon. including patient education. as it is a complex intervention. but outside of demonstration sites and experimental studies (Layard. 2000). 123 . Resource constraints have often been a significant limitation of these developments. Van Straten et al. it should be noted that the evidence base is largely from the US and. 1999). and medication management ● the provison of long-term coordination of care and follow-up. This work has been influenced by organisational developments in healthcare in the US. It can be summarised as follows: ● the provision of case management. for example the varying developments of the attached professional role over the past 20 years (Bower & Sibbald. and coordinated at either the primary or secondary care level. 2006. 2006) to call for the widespread implementation of collaborative care.... 2000). which is supervised and supported by a senior mental health professional ● the development of a close collaboration between primary and secondary care services ● the provison of a range of interventions consistent with those recommended in this guideline.

No additional studies were found for the attached professional models. The increased focus on social inclusion and the role of employment in maintaining good mental health led the GDG to also consider an updated review of employment but as no new studies were identified in the searches undertaken for this guideline the GDG decided not to update the review undertaken for the previous guideline. 1997) and computer programmes (Proudfoot et al. The use of these materials may be entirely patient managed.1 STEPPED CARE Introduction Stepped care seeks to identify the least restrictive and least costly intervention that will be effective for a person’s presenting problems (Davison. 2004). It is usually delivered by a pharmacist or nurse.1 with the exception of medication management. crisis resolution and home treatment teams and day hospitals were not updated. 2000). This was because they were the focus of considerable interest in the NHS and in the case of collaborative care considerable new evidence has emerged since the publication of the previous guideline. collaborative care. Medication management Medication management (for example. For similar reasons the reviews of social support systems. The GDG also decided to review medication management because there was evidence of increased use of this intervention in depression but considerable uncertainty as to whether the evidence supported medication management as a single. These may include self-help materials such as books (Cuijpers. 1999) is an intervention aimed at improving patient adherence to medication. stand-alone intervention. the attached professional model and medication management. It involves patient education about the nature and treatment of depression. the monitoring of side effects and the promotion of treatment adherence. or it may involve some limited input from a professional or paraprofessional. which is often 124 .4 5. the delivery of medication adherence strategies. which is given below..3. particularly in relation to collaborative care. Peveler et al. so the GDG decided that rather than performing a separate review they would comment on it.3. Definitions The definitions adopted are as stated in Section 5. which is often referred to as pure self-help. The lowintensity interventions most often used are those that are less dependent on the availability of professional staff and focus on patient-initiated approaches to treatment.3 Interventions included The GDG considered the range of interventions described above and the extent of current practice and decided to focus the reviews for this update on the following interventions: stepped care (including where possible matched care)..Case identification and service delivery 5. 5.4.

2 Databases searched and the inclusion/exclusion criteria The review team conducted a new systematic search for studies of stepped care in depression. 2007). each study considered for review is referred to by a ‘study ID’ made up of first author and publication date (unless a study is in press or only submitted for publication. consideration should be given to not only the degree of restrictiveness associated with a treatment and its costs and effectiveness.. but a stepped-care system is an explicit attempt to formalise the delivery and monitoring of patient flows through the system. Details of the search strings used are in Appendix 8. CINAHL Database inception to January 2008 July 2008. 125 . However.3 Studies considered20 The systematic review identified only one high-quality study (VANSTRATEN2006). 5. Escalating levels of response to the complexity or severity of the disorder are often implicit in the organisation and delivery of many healthcare interventions. 20Here and elsewhere in the guideline. ICD or similar criteria Stepped care referred to as guided self-help (Gellatly et al. This was undertaken as a joint review for this guideline and the guideline for depression in adults with a chronic physical health problem (NCCMH. 5. when first author only is used). January 2009 RCT People with a diagnosis of depression according to DSM. PsycINFO.4. this study included a sample of mixed depression and anxiety disorders and it was therefore decided to conduct a narrative review. but also the likelihood of its uptake by a patient and the likely impact that an unsuccessful intervention will have on the probability of other interventions being taken up. EMBASE.Case identification and service delivery Table 11: Databases searched and inclusion/exclusion criteria for clinical effectiveness of stepped care Electronic databases Date searched Update searches Study design Population Treatments MEDLINE.4. Information about the databases searched and the inclusion/exclusion criteria used are presented in Table 11. In establishing a stepped-care approach. 2010). which is set out below.

Other evidence suggests that individuals in stepped-care programmes may seek treatment in addition to the low-intensity interventions offered in the study and thereby undermine the efficiency assumption (Treasure et al.4 Narrative review In the field of mental health in the UK. with some suggestion that steppedcare models may be associated with lower rates of entry into studies (Whitfield et al. Thiels et al.. 2008) and the use of written materials (Cuijpers. the previous guideline on depression [NICE. from work on CCBT (Proudfoot et al. despite its widespread adoption. Bower and Gilbody (2005a) reviewed the evidence for the use of stepped care in the provision of psychological therapies and were unable to identify a significant body of evidence. The matched care control involved patients being allocated to an appropriate psychological treatment as determined by the responsible clinician. but outcomes were no better than for those patients in the other two arms. Patients in the matched control received more treatment sessions. For the efficiency assumption. stepped-care models are increasingly common and underpin the organisation and delivery of care in a number of recent NICE mental health guidelines (see for example. They set out three assumptions on which they argue a steppedcare framework should be built and which need to be considered in any evaluation of stepped care.. Bower and Gilbody (2005a) suggest that some of these problems could be addressed by taking into account patient choice (possibly by offering a choice from a range of low-intensity interventions) and also by adjusting the entry level into the stepped-care system to take account of the severity of the disorder.. Marks et al...4. for example. at least for some patients). 2002)..Case identification and service delivery 5. 1997). 2004b]). Some evidence for the equivalence of low-intensity interventions comes. 2001. unlike the other two arms of the trial where the type and duration of treatment were determined by the trial protocol. However. 1998). 1996.. this compared two forms of stepped care with a ‘matched care’ control. Since the publication of the Bower and Gilbody (2005a) review. They reviewed the existing evidence for stepped care against these three assumptions and found some evidence to suggest that stepped care may be a clinically and cost-effective system for the delivery of psychological therapies. 2003). the efficient use of resources (including healthcare resources outside the immediate provision of stepped care) and the acceptability of low-intensity interventions (to both patients and professionals). although there is some suggestion that CCBT may be more cost effective than therapist-delivered care (Kaltenthaler et al. but no evidence that strongly supported the overall effectiveness of the model. 2004. there is a limited evidence-base of studies designed specifically to evaluate stepped care. Kaltenthaler et al. 2004a] and the guideline on anxiety [NICE. Past experience of treatment or treatment failure may also be a useful indicator of which level a patient should be entered into the stepped-care model. These assumptions concern the equivalence of clinical outcomes (between minimal and more intensive interventions. Both forms of stepped care involved assignment to a psychological therapy. Further problems emerge when the acceptability assumption is considered. Although the study lacked 126 . brief behaviour therapy with a strong self-help component and therapist-delivered CBT. evidence is more difficult to identify. a study of stepped care for over 720 patients by Van Straten and colleagues (2006) has been published.

into additional treatments other than those delivered in the study.. the prescription of antidepressant drugs has been the first intervention offered (Katon et al. Evidence related to stepped care also comes from the Sequenced Treatment Alternatives to Relieve Depression (STAR*D) study (Rush et al.. Swindle et al. the efficacy of different antidepressants). augmentation treatments and. The study was designed to be as analogous as possible to real clinical practice. especially in the behaviour therapy group. This was a four-level study designed to assess treatments in patients who had not responded to previous treatment. but whether this reflects the specific contribution of a stepped-care framework is unclear. In order to achieve this.. both stepped-care arms had higher attrition rates and there was some diversion. 2003).. 1999. However. the IMPACT study [Unutzer et al. (A fuller review of the collaborative care literature is contained in Section 5. At each level. 2001... greater duration of intervention and follow-up and a greater range of available interventions. 2006). patients who had not responded to treatment at the previous level were randomised to different treatment options (or ‘stepped up’). patients were allowed to opt out of being randomised to drug switching. As may be apparent from this discussion. few of the studies have been built exclusively on a stepped-care model with all individuals receiving a low-intensity intervention at first point of contact. In many collaborative care studies.. 2002]) tend to be associated with better outcomes. to CBT. Clark 127 . They were not allowed to opt out of randomisation to a particular agent within the drug switching or drug augmentation arms. In addition. as such. it is possible that the two stepped-care models were more cost effective (Hakkaart-van Roijen et al.5. meta-regression studies such as those by Bower and colleagues (2006) and Gilbody and colleagues (2006) did not identify the presence of stepped care or specific algorithms of care (which may be taken as a rough equivalent or proxy for stepped care) as being associated with a more positive outcome. it can be said to be a form of stepped care. a number of other factors including the role of case management and other healthcare interventions may have an influence on the outcome.) Specifically in relation to collaborative care. in level two. Unutzer et al. considerable use has been made of stepped care programmes most notably in collaborative care (for example. but it did demonstrate that patients gained some benefit from moving through sequenced or stepped care and that it was possible to investigate this empirically. It is also the case that more complex collaborative care interventions (for example. The final evidence for the effectiveness of a stepped-care model in mental healthcare comes from the report on the two IAPT demonstration sites (for example.Case identification and service delivery power to determine whether the difference was statistically significant (despite including over 700 patients). A more limited number of studies have offered psychological interventions as the first point of contact (or the option of a pharmacological or psychological first treatment) in a collaborative care programme (Rost et al. 2006) where stepped care is often integrated into an overall collaborative care programme. 2002) and where benefit has not been obtained have stepped up either to more intensive pharmacological or psychological treatments or a combination of both. Outside the area of stepped care for psychological therapies for depression. 2003). for example. Hunkeler et al. The trial did not provide clear evidence on the suitable sequencing of treatment options (in particular.. The decision whether to step up to another intervention was then based on no or limited response to treatment.

the stepped-care model remains the best developed system for ensuring access to cost-effective interventions for a wide range of people with depression. 5..6. Since that time there has been further but limited evidence providing direct support for the model (Hakkaart-van Roijen et al. both of which provide a psychological stepped-care programme. the adoption of stepped-care models in non-mental healthcare has been associated with better physical health outcomes. Beyond the area of common mental health problems in fields such as addiction (Davison. 2000). a number of studies of stepped care have been conducted.4. there is some evidence for the effectiveness of the model.5 Health economic evidence and considerations No evidence on the cost effectiveness of the stepped care approach was identified by the systematic search of the economic literature. In the view of the GDG. Finally. It has also been adopted by the IAPT programme (Department of Health. 2006.. there is limited evidence from direct studies in common mental health problems that provide evidence for the effectiveness of the stepped-care model. while at the same time the outcomes obtained were broadly in line with those reported in RCTs for depression and anxiety. 2004b]). Bower and Gilbody (2005a) also provide some limited evidence in favour of the model in psychological therapies but. There is some evidence that the integration of stepped care into a more complex model of collaborative care may be associated with better outcomes but there is no direct evidence that this is the case. The 128 . particularly if supported by systems for routine outcome monitoring.1. 2008). 1999). Section 3. Outside the area of mental health. Clark et al. 2006. In the demonstration projects there was good evidence for increased patient flows through the system. 2007) as the framework for the delivery of the service. In summary.. no formal trials of the relative efficiency or effectiveness of a pure stepped-care model were identified.Case identification and service delivery et al. 2008) along with its increasing use in a number of collaborative care interventions. 5.. In each case there has been a positive benefit associated with stepped care. with the exception of the Van Straten and colleagues’ (2006) study. which ensure that there are systems in place that enable prompt stepping up for those who have not benefited from a low-intensity intervention. 2005) and acutely injured trauma survivors.. Van Straten et al. These include studies of stepped care in back pain (Von Korff.4.6 From evidence to recommendations The previous guideline recommended the adoption of a stepped-care model for the provision of psychological and pharmacological interventions for depression (the model was also used in the NICE guideline on anxiety [NICE. obesity (Carels et al. The adoption of the stepped-care model within the IAPT pilot sites was associated with the efficient use of healthcare resources and outcomes equivalent to those seen in clinical trials. Details on the methods used for the systematic search of the economic literature are described in Chapter 3.

Therefore.. In the UK. Rost et al.. Gilbody et al. Richards & Suckling. Other related reviews have focused on the use of case management in depression (Gensichen et al. Whittington et al..Case identification and service delivery GDG endorsed the model set out in the previous guideline but made some adjustments to the structure and content of the model in light of changes in the recommendations in this guideline update. 2008. Unutzer et al. Further research however is clearly needed to address the issues of efficacy. efficiency and acceptability of stepped care for depression. Pilling et al. Richards & Suckling. more recently. mental health professionals provided the enhanced staff input to primary care settings and undertook a care co-ordinator role (Katon et al. 2010). but not necessarily the cost effectiveness (Ofman et al. 2003) who may potentially affect this situation. there is a concern that there are not sufficient mental health professionals to provide enhanced input and care co-ordination for all primary care patients with depression. 2000) or graduates without core mental health professional training (Katzelnick et al.. The advent of these posts has recently been superseded by the development of the IAPT programme.. 1998. The model is set out in Figure 1 in Section 2. 2000. including primary care nurses (Mann et al.. 2006). Primary care nurses have multiple and increasing demands on their time. 2004. 1998). 2000.4.. one study used practice nurses in the care co-ordinator role and this did not improve either patient antidepressant uptake or outcomes compared with usual GP care (Mann et al. 2006).5 5.. and many are also uninterested in working with patients with psychological problems (Nolan et al.. 2004. 2006. 2002). Neumeyer-Gromen et al.. Current models are in development (for example.. 2008) that will allow service delivery systems to monitor and review the effectiveness of stepped-care models. Further elements were 129 . A number of recent meta-analyses of collaborative care have supported the statistical and clinical effectiveness of the model for depression (Badamgarav et al..7.. Simon et al. 2003. In many of the earlier studies.. 1995. which they defined as ‘an intervention for continuity of care including at least the systematic monitoring of symptoms. more recent studies have used mental health professionals or paraprofessionals (Chew-Graham et al. 1999). Most studies have been from the US. 2007. it seems unlikely that practice nurses will take on a significant role in the routine care of patients with depression.. where the role of lowintensity staff (in many cases a development of the primary care mental health worker role) has elements that are common to a number of collaborative care interventions. However. Katon et al. 2009). Department of Health... 1996.. 2000.1 COLLABORATIVE CARE Introduction The origins of collaborative care for depression lie in concerns about the inadequacy of much current treatment for the condition and developments in the field of chronic physical disorders.. 5.5. others. Gilbody et al. 2000). have taken on this role. Hunkeler et al. A major NHS staffing initiative for primary care mental health was the appointment of new graduate primary care mental health workers (Department of Health. In the UK.

3 and. 5.. 5. 1999).5. then a date is not used). References for studies from the previous guideline are in Appendix 18.70.. Of these. The effect sizes on depressive and related symptoms described in the reviews by Badamgarav and colleagues (2003). Definition This is set out in Section 5.25 (95% CI: 0.5. 28 were included and 22 were excluded. 130 .3 Studies considered21 In total. 2005).Case identification and service delivery possible such as coordination and assessment of treatment and arrangement of referrals’.81) (Neumeyer-Gromen et al. care management (Gensichen et al. the formal provision of collaborative care is not very evident in the NHS. 2006) and signposting (Grayer & Rudge. Current practice The extent of NHS-based provision in the UK has already been reviewed in Section 5. ranging between 0. The review by Whittington and colleagues (2009) is the only review that attempts to compare the effectiveness of collaborative care with the effectiveness of the attached professional model (Bower & Sibbald. Given this rather broad definition. Study IDs in title case refer to studies included in the previous guideline and study IDs in capital letters refer to studies found and included in this guideline update. 0.75 (95% CI: 0.. 2004). each study considered for review is referred to by a study ID (primary author and date of study publication.3. with most reviews reporting effect sizes at the lower end of the range indicated. 2006) and 0. 2010). Of the included studies. Neumeyer-Gromen and colleagues (2004).18.. although some elements of it are becoming available through the low-intensity arm of the IAPT programme. This was undertaken as a joint review for this guideline update and the guideline on depression in adults with a chronic physical health problem (NCCMH. except where a study is in press or only submitted for publication. 50 trials were found from searches of electronic databases. Gilbody and colleagues (2006) and Whittington and colleagues (2009) were generally modest. 11 were from the previous 21Here and elsewhere in the guideline. Information about the databases searched and the inclusion/exclusion criteria used are presented in Table 12. 2000). Details of the search strings used are in Appendix 8.32) (Gilbody et al. particularly in light of the considerable variation in the duration and complexity of the interventions covered in the meta-analyses described above.2 Databases searched and the inclusion/exclusion criteria The review team conducted a new systematic search for studies of collaborative care of depression. as can be seen from that section. including medication management (Peveler et al. the GDG did not consider that a separate analysis of case management from collaborative care was meaningful. 0.

Similarly. EMBASE. based in Chile.. Unutzer2002 was removed because of the high percentage of patients with chronic health problems reported in the study sample and this led the GDG to decide that the trial was more appropriately placed in the guideline on depression in adults with a chronic physical health problem (NCCMH. because it was identified as an outlier producing a great deal of heterogeneity (non-response data pre-sensitivity analysis I 2 82. post-sensitivity analysis I 2 0%). which was eradicated after it was taken out of the analysis (mean endpoint pre-sensitivity analysis I 2 43. These included the SCL-20 and SCL-depression subscale which are both depression-specific scales derived from the 90-item Hopkins Symptom Checklist (HSCL.Case identification and service delivery Table 12: Databases searched and inclusion/exclusion criteria for clinical effectiveness of collaborative care Electronic databases Date searched Update searches Study design Population Treatments MEDLINE. 2010). ICD or similar criteria Collaborative care. which was utilised as the control. case management. BDI-II (Beck et al. 1960).. one had been included in the previous guideline but was removed for the update because only 21% had a diagnosis of depression at baseline.6%. The GDG felt that this was a likely consequence of the study setting. the BDI (Beck et al. Derogatis. A range of self-rated and clinician-rated outcomes were reported in the included studies. The most common reasons for exclusion were that there was no extractable data or that less than 80% of participants had a diagnosis of depression.. Of the included studies. Wells1999 reported follow-up data at 45 months after the acute phase. Hamilton. which was not extracted because it was felt that the data could not be converted reliably into intention-to-treat analysis given the high attrition rate at that time point. feedback guideline (NCCMH. 1961). post-sensitivity analysis I 2 69. it is possible that the usual care arm. PsycINFO. All studies of populations with depression and an identified physical health problem (for example. One study reported follow-up relapse 131 . January 2009 RCT People with a diagnosis of depression according to DSM. CES-D (Radloff. CINAHL Database inception to January 2008 July 2008. 1974).2%). Araya2003 was also removed in a sensitivity analysis. 1977) and the Hamilton Rating Scale for Depression (HRSD. the major depression subsample from Katon1996MAJOR was removed from mean endpoint analysis because it too introduced an exceptionally large amount of heterogeneity. reflected a different healthcare system not relevant to a UK setting. PHQ-9 (Spitzer et al. 1996).2%. monitoring. 1999). 2004) and of the excluded studies. KATON2004) were excluded at the outset.

Appendix 16a for the full profile and Appendix 19a for the forest plots. thus strengthening the robustness of the original analysis. See Table 14 for the summary evidence profile. Where papers reported the intracluster correlation coefficient (ICC) this was used in the calculations. which also includes details of excluded studies23. Data were only extracted where a comparison with usual care was available. In these trials the unit of randomisation was the individual physician. The studies that were identified by the search and included in this review varied considerably in terms of the complexity of the care protocols implemented. although the effect sizes were small. antidepressant use.Case identification and service delivery prevention data. where k indicates the number of clusters.02 used where the ICC was not reported. In addition to this. There was also variation in participant diagnoses. 23For this review studies from the previous guideline were re-entered into the study database for the guideline update in Appendix 17a. The presence of such elements raises questions about the complexity or comprehensiveness of the intervention in particular when assessed against the criteria originally developed by Wagner (1996). with the empirically derived value of 0. 132 . 5. m the number of observations per cluster and ICC the intracluster correlation coefficient. The results indicated that applying the transformation had little to no impact on any of the results reported. Previous meta-regression had identified a number of factors such as mental health background of the care coordinator. A sensitivity analysis was conducted to compare the results of the metaanalysis with and without the application of the design effect. DIETRICH2004. Rost2001b. SWINDLE2003. and the provision of supervision as associated with better outcomes. DOBSCHA2006. collaborative care was more effective than standard care.5. and Wells1999). a number of trials employed a cluster randomised design. 22N (effective) = (k m)/(1 (m 1)) * ICC. A design effect22 was applied to the analysis of studies that had not accounted for the clustering in their analysis. Summary study characteristics of the included studies are in Table 13 with full details in Appendix 17a. studies including patients presenting with an antidepressant prescription were included along with those reporting a more formal diagnosis. In order to reduce the possible confounding crossover effects in which the implementation of collaborative care changes the standard care for all patients in the practice.4 Clinical evidence On some key outcome measures of efficacy such as self-rated non-response or mean end point scores. With this in mind a simple checklist (see Appendix 10) to assess the complexity of the intervention provided was used to see if this would help in more reliably characterising the interventions and ascertaining whether or not this would relate to the outcome of the intervention. healthcare firm or geographical area (DATTO2003. ROST20001a. the inclusion of both UK and non-UK based trials resulted in inevitable variation in the nature of the usual care used as a comparator. clinic.

191) (total participants) Study IDs (1) ADLER2004 (2) Araya2003 (3) Blanchard1995 (4) CHEWGRAHAM2007 (5) DATTO2003 (6) DIETRICH2004 (7) DOBSCHA2006 (8) FINLEY2003 (9) Hunkeler2000* (10) Katon1995 (11) Katon1996 (12) Katon1999 (13) Katon2001‡ (14) LUDMAN2007† (15) Mann1998b (1) 364/72 (2) 240/100 (3) 82/85 (4) 76/72 (5) 37/61 (6) 325/80 (7) 26/7 (8) 106/85 (9) 210/69 (10) 166/78 (11) 113/75 (12) 170/75 (13) 286/74 (14) 74/71 (15) Unclear for study 2 only (1) 42 (2) 43 (16) MCMAHON2007 (17) PERAHIA2008 (18) PILLING2010 (19) RICHARDS2008 (20) RICKLES2005 (21) ROST2001a24 (22) Rost2001b (23) Simon2000* (24) SIMON2004* (25) SIMON2006 (26) SMIT2006† (27) SWINDLE2003 (28) Unutzer2002 (29) Wells1999* (16) Unclear (17) 617/64 (18) 52/60 (19) 88/77 (20) 53/84 (21) Unclear for ‘recently treated’ only (22) 177/84 (23) 439/72 (24) 446/75 (25) 134/65 (26) 168/63 (27) 9/97 (28) 1168/65 (29) 981/72 (16) Unclear (17) 46 Continued N/% female Mean age (years) 24Presents acute data of Rost2001b. trials 28 RCTs (10. 133 . Rost2001b only used in analysis to avoid double counting.Case identification and service delivery Table 13: Summary study characteristics of collaborative care Collaborative care versus usual care No.

dysthymia or double depression (DSM–IV) (7) Subthreshold depressive symptoms. dysthymia or double depression (DSM–IV) (2) MDD (DSM–IV) (3) Probable pervasive depression (Short-CARE) (4) Unclear (5) MDD (MINI) or referred with depressive symptoms (6) MDD. dysthymia or double depression (DSM–IV) Continued 134 . dysthymia (DSM–IV) or unclear (8) Unclear: clinical judgment (9) MDD or dysthymia (DSM–IV) (10) MDD or subthreshold depressive symptoms (DSM–III-R) (11) MDD or subthreshold depressive symptoms (DSM–III-R) (18) 46 (19) 42 (20) 38 (21) Unclear for ‘recently treated’ only (22) 43 (23) 47 (24) 45 (25) 43 (26) 43 (27) 56 (28) 71 (29) 43 (16) Depressive illness (ICD–10) (17) MDD (DSM–IV) (18) Clinical diagnosis established by GP (unclear) (19) MDD (DSM–IV) (20) Unclear: antidepressant prescription (21) MDD (DSM–III-R) (22) MDD (DSM–III-R) (23) Unclear: antidepressant prescription (24) Unclear: beginning antidepressant treatment (25) Depressive disorder (unclear) (26) MDD (DSM–IV) (27) MDD. dysthymia. partially remitted MDD or double depression (PRIME-MD) (28) MDD.Case identification and service delivery Table 13: (Continued) Collaborative care versus usual care (3) 76 (4) 76 (5) 37 (6) 42 (7) 57 (8) 54 (9) 55 (10) 47 (11) 46 (12) 47 (13) 46 (14) 50 (15) Unclear for study 2 only Diagnosis (1) MDD.

†4-armed trial. ‡Relapse prevention study. double depression or subthreshold depression (CIDI) (17) Europe (18)–(19) UK (20)–(25) US (26) Netherlands (27)–(29) US (16) 180 (17) 84 (18) 120 (19) 90 (20) 90 (21) 730 (22) 730 (23) 112 (24) 180 (25) 84 (26) 1095 (27) 90 (28) 365 (29) 180 (18) 4 months (19)–(26) Not reported (27) 9 months (28) 6 and 12 months (29) 6 months Length of treatment (days) Follow-up *3-armed trial. dysthymic disorder.Case identification and service delivery Table 13: (Continued) Collaborative care versus usual care (12) Recurrent depression or dysthymia (DSM–IV) (13) Recovered but high risk of relapse (14) Subthreshold depressive symptoms or dysthymia (treatment resistant. 135 . DSM–IV) (15) MDD (DSM–III) Setting (1) US (2) Chile (3)–(4) UK (5)–(14) US (15)–(16) UK (1) 180 (2) 84 (3) 90 (4) 84 (5) 112 (6) 180 (7) 365 (8) 170 (9) 180 (10) 210 (11) 210 (12) 90 (13) 365 (14) 365 (15) 120 (1) 6 and 12 months (2) 3 months (3)–(10) Not reported (11) 7 months (12) 25 months (13)–(17) Not reported (29) MDD.

To estimate the potential effect of the attached professional role.46. However. SMD 0. One study. 0. The effect size for collaborative care in the review for this guideline was: self-rated outcome. 5. 136 .25. Given the similarity of effect sizes between the two modes of delivery of care. all trials for high-intensity psychological interventions included in the guideline were reviewed.35 (SMD. in some cases it involved case managers taking on the long-term care of people with depression (for example. This arose from a concern that a significant proportion of the data for the effectiveness of collaborative care was drawn from studies conducted in North America. Scott1997. These effect sizes were similar to that obtained in another review.5. The GDG did consider the inclusion of pharmacological trials based in primary care. Katon2001. it was decided to explore the potential effect sizes of the attached professional role versus usual GP care or waitlist control and therefore provide a comparator for collaborative care. The effect size for depressive symptomatology in that study was 0.28 (95% CI 0. Simpson2003 and Ward200025. 95% CI 0. The effect sizes for depressive symptoms obtained in the review for the attached professional role were: BDI. See Chapter 8 for further details.35 (95% CI 0. 2000). SMD 0. at least initially in the absence of any direct comparisons.5 Collaborative care: implications of data on the attached professional role As part of the collaborative care review.11). Hamilton Depression Rating Scale (HAMD). a number of factors need to be considered. Given the development of the attached professional role in primary care services in the UK (Bower & Sibbald. looked at relapse prevention in people who had achieved remission. that there may be little difference in effectiveness.06). SMD 0.58. it improved the number adhering to medication. 0. 0.16 (95% CI 0.66.10). 0. but because there were very few and collaborative care often involves antidepressant treatment as a minimum it was not felt to be a useful comparator. and the overlapping confidence intervals it seems reasonable to conclude. There was no difference between the number relapsing who had received collaborative care and the number relapsing who had received standard care. The following studies were identified and the study characteristics for these can be found in Chapter 8 and Appendix 17b: Schulberg1996. albeit one with somewhat different inclusion criteria. When attempting to understand these results. including the considerable variation in the nature of the collaborative care provided. the GDG wished to understand the potential impact of collaborative care for depression on the UK healthcare system.25). Simon 25<80% of participants met diagnosis of depression.Case identification and service delivery Receiving collaborative care appeared to make little difference to the number of people leaving treatment early. Scott1992. by Whittington and colleagues (2009).

n 1264 – – – – – RR 1.Table 14: Summary evidence profile for collaborative care versus standard care (acute-phase efficacy data) Clinician-rated RR 0. n 863 Service c-care 03.88 to 1.04) (29. n 1173 Service c-care 03.85 (0.7% versus 59.02 Service c-care 03.86 to 0.83 (0.4% versus 57.3%) High K 6.5% versus 29.97) (70% versus 76%) Quality High Number of studies.69 to 1. participants K Case identification and service delivery Forest plot number Service c-care 03.02 1820 K 2.86 (0.2%) Moderate K 1. n 1480 K 1.92) (49.06) (44.01 Non-remission RR 0.03 Continued Not reported DSM criteria Follow-up Not reported Self-rated Non-response RR 0.02 137 .91 (0.21) (49.2% versus 48.9 to 1. n 962 – RR 0.9%) Quality High Number of studies.98 (0.75 to 0.70 to 1. n Service c-care 03.05 (0.01 RR 0. participants K Forest plot number Service c-care 03.09) (56.7%) Moderate 7.5%) High 3.1% versus 47.

07 Self-rated Mean depression scores at endpoint SMD 0.06) Case identification and service delivery Quality High Number of studies. n 958 – – – – Not reported 1876 K 1.138 Table 14: (Continued) Clinician-rated SMD 0.36 ( 0. n 45 – – Not reported DSM criteria Follow-up 3–4 months SMD 0.05 ( 0.06 Not reported – – – Service c-care 03.53) High 11. n Service c-care 03.64 to 0. participants K Forest plot number Service c-care 03. participants – Forest plot number – .02 ( 0.25 to 0. n 214 Service c-care 03.15 to 0.11) Moderate K 1.63 to 0.16 ( 0.05 Mean depression change scores at endpoint Not reported Quality – Number of studies.05 SMD 0.09) High K 3.

for example. reporting a large effect on continuous data: SMD 1. for example Katon1996MAJOR. 5. It is also worth noting that when response data are reviewed. there were differences in the nature of the intervention provided – for example.92) and Katon1996MINOR: RR 0.97 (95% CI 0. 1997) and this may have had an impact on the effectiveness of the intervention. 139 . A review of collaborative care for this population.68 (95% CI 0. the majority of the attached professional studies were based in the UK (26 out of 38) and most of the collaborative care studies were based in the US (13 out of 16). the effect sizes for remission and response) and there is a small effect on endpoint continuous data. There was considerable variation between studies with some. In Whittington and colleagues’ (2009) review. Scott et al. in others it involved little more than advice and consultation with a psychiatrist (for example.. such as a meta-regression.64. 2010). and finally the comparators and the nature of the healthcare system in which the interventions were delivered may also have been different.Case identification and service delivery et al..7 Health economic evidence and considerations The systematic search of the economic literature undertaken for the guideline update identified no eligible studies on service-level interventions for people with depression set in the UK. 0. for example.. to more recent studies such as SIMON2006: RR 0.15).18). 0. but inclusion of this study in the meta-analysis resulted in considerable heterogeneity.5.1.59). The evidence profiles developed for this guideline show that when the review of collaborative care is restricted to the groups with depression and no significant chronic physical health problems. 5.5.41 1. 2004). which entirely disappeared when the study was removed in the sensitivity analysis.27. 1. there is a noticeable decline in effect size from the early studies for example. Section 3.6.6 Clinical summary The studies of collaborative care reviewed here were limited to people without an accompanying chronic physical health problem. then the effects of the intervention are of limited clinical importance (see.49 (95% CI 0. It should also be noted that the endpoint continuous data effect sizes were similar to those obtained from an analysis of the attached professional role. The small size of the dataset included here prevented any more detailed analysis. including studies of older people with a high incidence of physical health problems (Unutzer2001) is contained in the related guideline (NCCMH. Katon1996MAJOR: RR 0. Katon et al. the populations included in the trials may have been different. 1995). within the attached professional model. the professionals more consistently provided specific psychological interventions (for example. Details on the methods used for the systematic search of the economic literature are described in Chapter 3.11 (95% CI 1.79.

The GDG did recognise that co-ordinated care with long-term follow-up is an important element of effective care for people with severe and complex depression. In addition. NCCMH. 5.. The collaborative care meta-analysis conducted for the update points to a small effect size of collaborative care when compared with usual care. it is acknowledged in the guideline on depression in adults with a chronic physical health problem (NICE. the additional input of a case manager in the co-ordination of care for people with depression and associated liaison time with GPs and specialist psychiatrists. The effect sizes were considered too small to warrant a formal economic evaluation. Collaborative care studies included in the meta-analysis point to a resource use that is more intensive than usual care. A decision was reached by the GDG not to recommend collaborative care interventions in the depressed population in the absence of chronic physical health problems. for example.5. A significant portion of the effectiveness data was based on studies conducted in the US and collaborative care is a service-level intervention with effects that largely reflect the nature of the healthcare setting in which it is provided. co-ordinated multiprofessional interventions are key elements of the care provided in specialist mental health services in Step 4 of the stepped-care model in this guideline (see Figure 1 in Section 2. Therefore more studies conducted in the UK healthcare setting may provide more UK-specific effects and resource use estimates. The development of an approach to collaborative care for depression built on the provision of low-intensity interventions (such as behavioural activation and medication management) has shown promise in pilot trials in an NHS setting (for example. Richards et al. 2010) that collaborative care can play an important role in that population. In view of this. From this one can assume that collaborative care may be more costly than usual care. 2008) and the current multicentre Medical Research Council funded Collaborative Depression Trial (‘CADET’) may provide more substantial evidence for this type of intervention. 140 .4. which should inform further updates of this guideline. However this does not exclude the possibility of collaborative care being cost effective when compared with usual care as even small differences in effects and costs could potentially result in a cost-effective intervention.8 From evidence to recommendations The evidence reviewed in this guideline update for collaborative care in depression was not viewed as being sufficiently strong to generate any recommendations.Case identification and service delivery No UK-based studies evaluating the cost effectiveness of collaborative care were identified in the literature search. For example.7). the GDG thought it appropriate to draw attention in the recommendations to the role of collaborative care for people with depression and a chronic physical health problem and of co-ordinated multi-professional care in specialist mental health services for those with severe and complex depression. 2009c.

Dunn and colleagues (1999). in a study of over 16. 5. two conditions need to hold.6.000 primary care patients prescribed either TCAs or SSRIs. prescriptions covering at least 120 days’ treatment within the first 6 months after diagnosis) only 6% of those prescribed a TCA did. Teams working with people with complex and severe depression should develop comprehensive multidisciplinary care plans in collaboration with the person with depression (and their family or carer. First.5.5.2 5. if agreed with the person).5.9 5.9.1 Recommendations For people with severe depression and those with moderate depression and complex problems. NCCMH. The care plan should: ● identify clearly the roles and responsibilities of all health and social care professionals involved ● develop a crisis plan that identifies potential triggers that could lead to a crisis and strategies to manage such triggers ● be shared with the GP and the person with depression and other relevant people involved in the person’s care. If the potential benefits of longer-term treatment are to be realised. that the drugs are prescribed at an adequate dose and second that the regime of treatment is adhered to. 141 . evidence from studies of 26Refer to ‘Depression in adults with a chronic physical health problem: treatment and management’ (NICE. consider: ● referring to specialist mental health services for a programme of coordinated multiprofessional care ● providing collaborative care if the depression is in the context of a chronic physical health problem with associated functional impairment26. reported that while 33% of those prescribed an SSRI were judged to have completed an adequate period of treatment (that is. as has the problem of poor compliance.6 5.Case identification and service delivery 5. 2009c. inevitably this has stimulated interest in developing strategies to promote and support adherence to antidepressant medication. However.9. 2010) for the evidence base for this. Of course this study does not account for the possibility that some patients may have switched medication and may have done so to their long-term benefit.1 MEDICATION MANAGEMENT Introduction The effectiveness of antidepressants in the treatment of depression has long been recognised.

Information about the databases searched and the inclusion/exclusion criteria used are presented in Table 15. which suggests that collaborative care was associated with increased medication adherence. Lingam and Scott (2002) in a systematic review report non-adherence rates between 10 and 60% for antidepressants. NICE (2009b) has produced guidance on promoting medication adherence. However. report that only 35% ever received one prescription and only a minority received further prescriptions.6. Most recently. modest effects. then a date is not used). 0. 5.75).3 Studies considered27 In this guideline update four trials with potential relevance to medication management for depression were found from searches of electronic databases. This was undertaken as a joint review for this guideline and the guideline for depression in adults with a chronic physical health problem (NCCMH. Vergouwen and colleagues (2003) in a review of medication adherence compared interventions such as educational interventions not associated with a collaborative care intervention with those adherence programmes nested in collaborative care interventions. 5. Isacsson and colleagues (1999). and reported improved adherence and better clinical outcomes in the latter. 95% CI 0. in a study of nearly 1000 patients.58.Case identification and service delivery prescribing patterns in primary care suggests that if patients discontinue one form of antidepressant medication they often do not take another medication. at best. except where a study is in press or only submitted for publication. 142 . the effects of antidepressants seem modest and adherence to treatment regimes is also limited.2 Databases searched and the inclusion/exclusion criteria The review team conducted a new systematic search for studies of medication management. the problem of poor medication adherence has been the subject of considerable research and debate. the GDG was specifically concerned with the effectiveness of medication adherence (medication management programmes) in depression. 2010). with an average around 40%. They were also able to identify only a few well-conducted studies designed to improve antidepressant adherence with. For example. such as those developed by Katon and colleagues (2002). which has general applicability for promoting adherence across all fields of medical care. Details of the search strings used are in Appendix 8. which suggests a potentially positive impact on medication adherence (RR 0. This presents a potentially worrying picture.44.6. For example. Of these three 27Each study considered for review is referred to by a study ID in capital letters (primary author and date of study publication. This view of increased adherence to antidepressants in collaborative care was also supported by the metaregression study of Bower and colleagues (2006). Beyond depression and mental health. and by the review conducted for this guideline update of outcomes for collaborative care.

Summary study characteristics of the included studies are in Table 16 with full details in Appendix 17a. the overall outcomes were extracted for this review.5 Clinical evidence summary A total of five studies. In order to be consistent with the other studies. PsycINFO. CROCKETT2006 was a cluster randomised trial but the outcomes could not be adjusted because the number of clusters was not reported in the study. 2004a). EMBASE. See Table 17 for the summary evidence profile. 5. CINAHL Database inception to January 2008 July 2008. ICD or similar criteria Medication management were included and one was excluded because it did not report any outcomes relevant to the scope. which also includes details of excluded studies. It is therefore reported separately. PEVELER1999 reported both overall outcomes for all participants and an analysis of a subsample of more severely depressed patients. None of the studies was included in the previous guideline (NICE. were reviewed. the quality of the evidence from these five studies was 143 . Overall. focusing specifically on medication management in depression.4 Clinical evidence There was insufficient evidence that medication management helped to reduce symptoms of depression. 5. although it had some effect on medication adherence and appeared acceptable to participants.Case identification and service delivery Table 15: Databases searched and inclusion/exclusion criteria for clinical effectiveness of medication management Electronic databases Date searched Update searches Study design Population Treatments MEDLINE.6.6. January 2009 RCT People with a diagnosis of depression according to DSM. but it should be noted that the authors reported a significant effect for patients who met criteria for major depression at the outset and received TCAs at doses above 75 mg per day. Appendix 16a for the full profile and Appendix 19a for the forest plots. Of the included studies. The GDG identified two studies from the collaborative care review that were relevant to medication management so these were also included.

144 . clinical diagnosis) (4) Antidepressant prescription (unclear) (5) Depressive disorder (unclear) (1) US (2) Australia (3) UK (4) US (5) UK (1) 180 (2) 60 (3) 84 (4) 90 (5) 56 (1) 6 and 12 months (2)–(5) Not reported N/% female Mean age Diagnosis Setting Length of treatment (days) Follow-up *From the collaborative care review. trials (Total participants) Study IDs 5 RCTs (963) (1) ADLER2004* (2) CROCKETT2006 (3) PEVELER1999† (4) RICKLES2005* (5) WILKINSON1993 (1) 364/72 (2) 84/71 (3) 157/74 (4) 53/84 (5) 45/74 (1) 42 (2) 46 (3) 45 (4) 38 (5) 49 (1) MDD. dysthymia. double depression (DSM–IV) (2) Antidepressant prescription (unclear) (3) Depressive illness (unclear. †4-armed trial.Case identification and service delivery Table 16: Summary study characteristics of medication management versus usual care Medication management versus usual care No.

63 to 1. Vergouwen et al. but even allowing for this. n 594 Service med-man 02. 2003).47 to 1.4%) Low K 1.Case identification and service delivery Table 17: Summary evidence profile for medication management versus usual care Self-rated Non-response Quality Number of studies.2% versus 34.4%) Moderate K 2.46 to 1.14 ( 0.05) (25. and it was not possible to perform a single meta-analysis of all the studies.9%) High K 2. participants Forest plot number Mean depression scores at endpoint Quality Number of studies.81 (0. the GDG did not feel able to make any recommendations for medication management alone in the treatment of depression.01 limited.02) High K 3.8% versus 40.08) (32. Where possible data from studies was combined. participants Forest plot number Adherence Quality Number of studies. n 604 Service med-man 01. Where there are specific concerns about potential 145 .5% versus 31. This is consistent with other reviews in the area (for example.89) (32. participants Forest plot number RR 0.03 RR 0. focusing on depression outcomes. 2009b) are potentially important in improving adherence. n 221 Service med-man 01.01 SMD 0. participants Forest plot number Leaving treatment early for any reason (including lost to follow-up) Quality Number of studies. However. it is recognised that the recommendations set out in the NICE guideline on medicines adherence (NICE. n 63 Service med-man 01..02 RR 0.31 to 0. no consistent picture of a clinically important benefit of medication management alone emerged from the data.7 (0. In light of this.94 (0.

5. It is likely to be effective only when provided as part of a more complex intervention. Crisis resolution and home treatment teams (CRHTTs) are a form of service that aims to offer intensive home-based support in order to provide the best care for someone with depression where this is the most appropriate setting.6.Case identification and service delivery problems with adherence (for example increased side effects with TCAs.6 Health economic evidence and considerations No evidence on the cost effectiveness of medication management was identified by the systematic search of the economic literature.6.1 Introduction **Traditionally.6.7 From evidence to recommendations The evidence reviewed in this guideline for medication management alone was not viewed as being sufficiently strong to generate any positive recommendations.7. However. Details on the methods used for the systematic search of the economic literature are described in Chapter 3. 5. 10.1.1 Recommendations Medication management as a separate intervention for people with depression should not be provided routinely by services. it may avoid the stigma and costs associated with hospital admission.6. 5.6. 5. Section 3. thus providing benefits to both patients and service providers. If this can be done safely. in recent years there has been growing interest in attempting to manage such episodes in the community. 146 . the delay in onset of antidepressant effects or the possibility of discontinuation symptoms) specific attention is drawn to these within the recommendations on pharmacological interventions (in Chapters 9.7 CRISIS RESOLUTION AND HOME TREATMENT TEAMS The following sections marked by asterisks (**_**) are from the previous guideline and have not been updated for this guideline except for style and minor clarification. 11 and 12). a depressive episode marked by serious risk to self (most often suicidal ideation and intent) or very severe deterioration to care for the self is managed by admission to an acute inpatient unit.8 5.8. 5.

Characteristics of the included studies are in Appendix 17a. only Stein1975 met the inclusion criteria set by the GDG (all the other studies had a very significant or exclusive focus on schizophrenia).7. 147 . Crisis intervention and the comparator treatment were defined as follows: ● Crisis resolution is any type of crisis-oriented treatment of an acute psychiatric episode by staff with a specific remit to deal with such situations. A further search identified no new RCTs suitable for inclusion. Stein1975). The GDG for the guideline update preferred the term ‘clinically important’ to avoid confusion with the term ‘statistically significant’. 2003). which included five RCTs (Fenton1979. Studies were excluded if they were largely restricted to people who were under 18 years or over 65 years old.7. in and beyond ‘office hours’. The GDG chose to use the Cochrane review of CRHTTs (Joy et al. Of the five RCTs included in the Cochrane review.2 Studies considered28 The focus of this review is to examine the effects of CRHTT care for people with serious mental illness (where the majority of the sample was diagnosed with nonpsychotic disorders) experiencing an acute episode compared with the standard care they would normally receive. 5.. which also includes details of excluded studies.3 Clinical evidence statements29 Crisis resolution and home treatment teams versus standard care Effect of treatment on death (suicide or death in suspicious circumstances) There is insufficient evidence to determine whether there is a clinically important30 difference between CRHTTs and ‘standard care’ on reducing the likelihood of death 28The study IDs for studies that were in the previous guideline are in title case. 29The forest plots can be found in Appendix 19a. Pasamanick1964. as the starting point for this section. Hoult1981. Muijen21992. ● ‘Standard care’ is the normal care given to those experiencing acute psychiatric episodes in the area concerned. this involved hospital-based treatment for all studies included.Case identification and service delivery Definition The GDG adopted the definition of crisis resolution developed by the Cochrane review of crisis intervention for people with serious mental health problems (Joy et al. 5. References for these studies are in Appendix 18. 30Note that the wording in the previous guideline was ‘clinically significant’.. or to those with a primary diagnosis of substance misuse or organic brain disorder. providing data for 130 participants. 2003).

N 130.73 to 0. 95% CI.51 to 1. 0.4 Clinical summary The very large majority of patients with depression are never admitted to hospital (in contrast to schizophrenia where 60 to 70% are admitted to hospital at first presentation. Effect of treatment on burden to family life There is insufficient evidence to determine whether there is a clinically important difference between CRHTTs and ‘standard care’ on reducing the likelihood of a patient’s family reporting disruption to their daily routine due to the patient’s illness by 3 months (K 1.83. There is insufficient evidence to determine whether there is a clinically important difference between CRHTTs and ‘standard care’ on reducing the likelihood of a patient’s family reporting significant disruption to their social life due to the patient’s illness by 3 months (K 1. There is evidence suggesting that there is a statistically significant difference favouring CRHTTs over ‘standard care’ on reducing the likelihood of a patient’s family reporting physical illness due to the patient’s illness by 3 months but the size of this difference is unlikely to be of clinical importance (K 1. Effect of treatment on burden to community There is insufficient evidence to determine whether there is a clinically important difference between CRHTTs and ‘standard care’ on reducing the likelihood of patients being arrested (K 1. 0. 0. Therefore. N 130.51 to 1. 1999). N 130. RR 1.41) or by 20 months (K 1. RR 0. N 130. RR 0.60.10).17. 95% CI.7. RR 0. 95% CI.41 to 3.02). 95% CI. RR 1.15 to 2.45). 0.76. N 130.70 to 1. N 130. 0.65). 0. RR 0.84.67 to 1.12).Case identification and service delivery due to any cause taking place during the study (K CI.79. There is insufficient evidence to determine whether there is a clinically important difference between CRHTTs and ‘standard care’ on reducing the likelihood of patients using emergency services (K 1. 95% CI. it is unsurprising that much of the evidence base is drawn from the treatment of schizophrenia and this means that there is currently insufficient evidence from RCTs to determine the value of CRHTTs for 148 . N 130.28). RR 0. 0. RR 0.95). RR 0.88. 95% CI. 95% Effect of treatment on acceptability There is insufficient evidence to determine whether there is a clinically important difference between CRHTTs and ‘standard care’ on reducing the likelihood of patients leaving the study early by 6 or 12 months (K 1. 1. There is some evidence suggesting a clinically important difference favouring CRHTTs over ‘standard care’ on reducing the likelihood of a patient’s family reporting physical illness due to the patient’s illness by 6 months (K 1. 95% CI.96). 95% CI.86. 0.00. 5.06 to 15. 0. McGorry & Jackson.66 to 0. N 130. N 130.

consider increasing the intensity and duration of the interventions and ensure that they can be provided effectively and efficiently on discharge. 1990). 1989) and the UK (Dick et al.. 5.1 Introduction **Given the substantial costs and high level of use of inpatient care. self-harm or self-neglect33.Case identification and service delivery people with depression.7.7.1 Recommendations Use crisis resolution and home treatment teams to manage crises for people with severe depression who present significant risk. and to deliver high-quality acute care.** 5. the possibility of day hospital treatment programmes acting as an alternative to acute admission gained credence in the early 1960s. 1971) and later in Europe (Wiersma et al. Consider crisis resolution and home treatment teams for people with depression who might benefit from early discharge from hospital after a period of inpatient care32. 5.5.5. 1965.5 5.3 5. The teams should monitor risk as a high-priority routine activity in a way that allows people to continue their lives without disruption31. Any wording changes have been made for clarification only.. 33Ibid..8. Nevertheless.4 5.8 ACUTE DAY HOSPITAL CARE The following sections marked by asterisks (**_**) are from the previous guildeline and have not been updated for this guideline except for style and minor clarification. 149 . The full range of high-intensity psychological interventions should normally be offered in inpatient settings.. Herz et al.7.2 5.7.5. CRHTTs may have value for that small group of patients with depression who require a higher level of care than can be provided by standard community services. Creed et al. However. Definition Acute psychiatric day hospitals were defined for the purposes of the guideline as units that provide diagnostic and treatment services for acutely ill individuals who would 31The evidence for this recommendation has not been updated since the previous guideline. initially in the US (Kris. Consider inpatient treatment for people with depression who are at significant risk of suicide. 1985.7. 32Ibid.5.

Clearly. 150 .74 to 1. Participants were people with acute psychiatric disorders (where the majority of the sample were diagnosed with non-psychotic disorders) who would have been admitted to inpatient care had the acute day hospital not been available.8. 5.3 Clinical evidence statements35 The studies included in this review examined the use of acute day hospitals as an alternative to acute admission to an inpatient unit. 35The forest plots can be found in Appendix 19a. References for these studies are in Appendix 18. Sledge1996) met the inclusion criteria set by the GDG. Moreover. RR 1. and some studies explicitly excluded people with families unable to provide effective support at home. Characteristics of the included studies are in Appendix 17a.43). 2001) as the basis for this section. Studies were excluded if they were largely restricted to people who were under 18 years or over 65 years old.02. including between 50 and 62% of people with a diagnosis of mood or anxiety disorder. 0. which also includes details of excluded studies. cannot be generalised to all people with depression who present for acute admission. Effect of treatment on efficacy There is insufficient evidence to determine whether there is a clinically important difference between acute day hospitals and inpatient care on reducing the likelihood of readmission to hospital after discharge from treatment (K 2. or to those with a primary diagnosis of substance misuse or organic brain disorder. 95% CI. acute day hospitals are not suitable for people subject to compulsory treatment. trials were eligible for inclusion only if they compared admission to an acute day hospital with admission to an inpatient unit..8. N 288. Of the nine studies included in the existing review. 5. the findings from this review. Thus. 34Study IDs in title case refer to studies included in the previous guideline. Marshall and colleagues (2001) focused on adults up to the age of 65 and reviewed nine trials of acute day hospital treatment published between 1966 and 2000. The individuals involved in the studies were a diagnostically mixed group. providing data for 288 participants.2 Studies considered34 The GDG selected a Health Technology Assessment (Marshall et al. only two (Dick1985. and the recommendations based upon them.Case identification and service delivery otherwise be treated in traditional psychiatric inpatient units. A further search identified no new RCTs suitable for inclusion.

.9.76). 0. RR 0.29 to 2.Case identification and service delivery Effect of treatment on inpatient days per month There is some evidence suggesting that there is a clinically important difference favouring acute day hospitals over inpatient care on inpatient days per month (K 1.9 NON-ACUTE DAY HOSPITAL CARE The following sections marked by asterisks (**_**) are from the previous guideline and have not been updated for this guideline except for style and minor clarification. 95% CI.1 Introduction **Although the earliest use of day hospitals in mental healthcare was to provide an alternative to inpatient care (Cameron.** 5.2 Studies considered36 The GDG chose to use the Cochrane systematic review (Marshall et al.46 to –0. N 197. and those with more severe and enduring mental disorders such as schizophrenia. the review team undertook a review of the evidence comparing the efficacy of non-acute day hospitals with that of traditional outpatient treatment programmes. 151 . 95% CI. in so far as they apply to people with serious mental health problems: ● psychiatric day hospitals offering continuing care to people with severe mental disorders. Studies were excluded if the participants were predominantly either over 65 years or under 18 years of age. Two broad groups of people have been referred for non-acute day hospital care: those with anxiety and depressive disorders who have residual or persistent symptoms. the GDG agreed the following definition for non-acute day hospitals. References for these studies are in Appendix 18.9. 5.59). –3. 5. WMD –2. Effect of treatment on acceptability There is insufficient evidence to determine whether there is a clinically important difference between acute day hospitals and inpatient care on reducing the likelihood of patients leaving the study early for any reason (K 2. Given the need for services for people with severe and enduring mental health problems that are refractory to other forms of treatment. 2003) that compared day treatment programmes with outpatient care for people with 36Study IDs in title case refer to studies included in the previous guideline.86. non-acute day hospitals have also been used for people with refractory mental health problems unresponsive to treatment in outpatient clinics.11. 1947). N 288. Definition For this section.

Effect of treatment on acceptability There is insufficient evidence to determine whether there is a clinically important difference between non-acute day hospitals and outpatient care on reducing the likelihood of patients reporting that they were not satisfied with care (assuming that people who left early were dissatisfied. WMD –3. Piper1993.42. Effect of treatment on efficacy There is insufficient evidence to determine whether there is a clinically important difference between non-acute day hospitals and outpatient care on reducing the likelihood of admission to hospital during the study at 6 to 8 months (K 2.97. 0. 95% CI. Piper1993. Characteristics of the included studies are in Appendix 17a.05).19).81. There is insufficient evidence to determine whether there is a clinically important difference between non-acute day hospitals and outpatient care on improving the patient’s mental state (change from baseline on the Present State Examination [PSE]) at 4 months (K 1. 95% CI. 37The forest plots can be found in Appendix 19a.9. –8. Dick1991. WMD –3. 95% CI. N 89. 0. RR 0.Case identification and service delivery non-psychotic disorders.54 to 6.24.3 Clinical evidence statements37 Effect of treatment on death (all causes) There is insufficient evidence to determine whether there is a clinically important difference between non-acute day hospitals and outpatient care on reducing the likelihood of death during the study (K 1.68 to 1. 0. RR 1. –8. which also includes details of excluded studies. Tyrer1979). 95% CI. WMD –4.39). N 89. Effect of treatment on social functioning There is insufficient evidence to determine whether there is a clinically important difference between non-acute day hospitals and outpatient care on improving the patient’s social functioning (change from baseline on the Social Functioning Schedule [SFS]) at 4 months (K 1. 95% CI. 5. N 89. Tyrer1979) were included providing data on 428 participants.25) and at 8 months (K 1.07 to 1.85). K 2. 95% CI. Bateman1999 was excluded from the review for this guideline because the sample were patients diagnosed with borderline personality disorder. three studies (Dick1991.38 to 5. WMD –3. N 88. –8.76) and at 24 months (K 1.95 to 1. as the starting point for this section. 95% CI. N 106.59) and at 8 months (K 1. RR 2. N 106. N 202. 0. –9. RR 1.48.18).72. 95% CI. Therefore.69 to 1.39. 152 .96 to 2.23 to 25.38. Of the four studies included in the Cochrane review (Bateman1999. N 200.

Although such self-help groups are likely to be beneficial. 1996) in a variety of nonhealthcare settings may confer some benefit and it is hoped that such projects are the subject of more formal evaluation.10.07). at about 12 months (K 1. 2000). the Newpin Project.61.35.9. Given that social isolation is associated with poor outcome and chronicity in depression. 0. 0. Several descriptive reports suggest that the provision of social support (for example. 95% CI. Despite this and the considerable amount of work that has described the importance of social support. 95% CI. 0. 153 .Case identification and service delivery There is insufficient evidence to determine whether there is a clinically important difference between non-acute day hospitals and outpatient care on reducing the number of people lost to follow-up at 6 to 8 months (K 2. N 106.1 Introduction **It is widely accepted that social support can play an important part in a person’s propensity to develop depression and their ability to recover from it. mental state. few formal studies of the potential therapeutic benefits of different forms of social support have been undertaken.49 to 2. There is also evidence to suggest that supported engagement with a range of non-statutory sector services is beneficial. 1978).. this is regrettable. death. N 202. Mills & Pound. 5.85 to 3. 5. There are many organisations offering local group peer support to people with depression. N 226.38). but this study was not limited to patients with depression and so was excluded from the review (Grant et al. social functioning or acceptability of treatment. including Depression Alliance and Mind. the review team were unable to find any research evidence for their effectiveness.4 Clinical summary There is currently insufficient evidence to determine whether acute day hospital care differs from inpatient care in terms of readmission to hospital after discharge. 95% CI.10 NON-STATUTORY SUPPORT The following sections marked by asterisks (**_**) are from the previous guideline and have not been updated for this guideline except for style and minor clarification. the evidence is inconclusive although there is a trend favouring day hospitals.08.94) and at 24 months (K 1. RR 1. There is evidence from a series of studies that providing social support in the sense of befriending (women with depression) confers benefits (Brown & Harris. RR 1. With regard to treatment acceptability.** 5.94 to 1. There is currently insufficient evidence to determine whether non-acute day hospital care differs from outpatient care in terms of admission to hospital. RR 1.

58. this did not reach statistical significance. Befriendees were women with chronic depression in inner London who were interested in being befriended. 154 . to broaden their range of activities. 40Depressed mood at four out of 10 symptoms on the PSE-10. 5. which also includes details of excluded studies. as opposed to 2 to 6 months. In this trial befriending was defined as ‘meeting and talking with a depressed woman for a minimum of one hour each week and acting as a friend to her. Women were allowed to be on other treatments such as antidepressants and contact with other healthcare professionals. References for these studies are in Appendix 18. Although remission tended to be higher among those completing the full 12 months of befriending. 0. Characteristics of the included study are in Appendix 17a. On an intention-to-treat analysis a clinically important effect upon remission was found at 1 year: There is some evidence suggesting that there is a clinically important difference favouring befriending over waitlist control on increasing the likelihood of patients achieving remission (defined as patients not meeting ‘caseness’ for depression40) (K 1. to offer practical support with ongoing difficulties and to help create ‘fresh start’ experiences often found to precede remission in previous work.93).10. N 86.36 to 0. Other treatments monitored naturalistically did not relate to remission nor did initial duration of chronic episode or comorbidity. so a descriptive review of the data is presented here.10. which provide support. 38Study IDs in title case refer to studies included in the previous guideline.Case identification and service delivery Definition The GDG agreed the following definition for non-statutory support: ● A range of community-based interventions often not provided by healthcare professionals. This suggests that the benefits of befriending may be obtained by a shorter intervention.2 Studies considered38 The review team found one RCT (Harris1999) of befriending compared with waitlist control in people with depression. listening and “being there for her”’.3 Clinical evidence statements39 Befriending versus wait list control One RCT of befriending (Harris1999) was identified. 5. activities and social contact in order to improve the outcome of depression. 95% CI. RR 0. 39The forest plots can be found in Appendix 19a. The trained volunteer female befrienders were also encouraged to accompany their ‘befriendee’ on trips.

1 RESEARCH RECOMMENDATION The efficacy and cost effectiveness of different systems for the organisation of care for people with depression.Case identification and service delivery Additional trials with less restricted intake conditions and in more naturalistic general practice settings might confirm volunteer befriending as a useful adjunct to current treatments. This question should be answered using a randomised controlled design which reports short-term and medium-term outcomes (including cost-effectiveness 41The evidence for this part of the recommendation has not been updated since the previous guideline. 155 .11. Stepped-care models are widely implemented but the efficacy of this model compared with matched care is uncertain.5 5.10. Any wording changes have been made for clarification only. befriending should be by trained volunteers providing. consider: ● befriending as an adjunct to pharmacological or psychological treatments.** 5. In people with mild. 5.10.1.1 Recommendation For people with long-standing moderate or severe depression who would benefit from additional social or vocational support. what system of care (stepped care versus matched care) is more clinically effective and cost effective in improving outcomes? Why this is important The best structures for the delivery of effective care for depression are poorly understood. The results of this study will have important implications for the structure of depression treatment services in the NHS. Evidence on the relative benefits of the two approaches and the differential effects by depression severity is needed. at least weekly contact for between 2 and 6 months41 ● a rehabilitation programme if a person’s depression has resulted in loss of work or disengagement from other social activities over a longer term.11 5. moderate or severe depression. 5.5. typically.10.4 Clinical summary There is some evidence that befriending given to women with chronic depression as an adjunct to drug or psychological treatment may increase the likelihood of remission.

In stepped care the majority of patients will first be offered a low-intensity intervention by a paraprofessional unless there are significant risk factors dictating otherwise. 156 . The full range of effective interventions (both psychological and pharmacological) should be made available in both arms of the trial. The study needs to be large enough to determine the presence or absence of clinically important effects. In matched care a comprehensive mental health assessment will determine which intervention a patient should receive. The outcomes chosen should reflect both observer and patient-rated assessments of improvement and an assessment of the acceptability of the treatment options.Case identification and service delivery outcomes) of at least 18 months’ duration. and moderators (including the severity of depression) of response should be investigated.

However. Since the previous NICE guideline on depression (NICE. Significant national initiatives are beginning to explore how to maximise the accessibility. This chapter sets out how these treatments have emerged as evidence-based approaches and some of the contextual issues that are important in translating recommendations based on clinical research to people presenting to the NHS with depression.Introduction to psychological and psychosocial interventions 6 INTRODUCTION TO PSYCHOLOGICAL AND PSYCHOSOCIAL INTERVENTIONS 6. there are sufficient developments to necessitate a significant review of the literature with consequent refinements to recommendations from the previous guideline.1 INTRODUCTION A range of psychological and psychosocial interventions for depression have been shown to relieve the symptoms of the condition and there is growing evidence that psychosocial and psychological therapies can help people recover from depression in the longer-term (NICE. 6. 2004a). This means that treatments need to have been shown to work against robust criteria that support evidence-based practice (see Chapter 3) and which are likely to be cost effective. acceptability and cost effectiveness of psychological and psychosocial interventions. However.. While a broad array of psychosocial and alternative therapies may be accessed by people seeking help with depression. Therefore there is a need to offer a range of psychological and psychosocial interventions and for further clinical innovation focused on improving treatment outcomes.2 RECOMMENDING PSYCHOLOGICAL AND PSYCHOSOCIAL TREATMENTS This guideline is concerned with promoting clinically and cost-effective treatments that should be provided in the NHS. if funded. It is important to note the limitations of the available data for making recommendations about psychological and psychosocial treatments (see Pilling [2008] for a fuller discussion of these issues). could inform the recommendations of future clinical guidelines. the extent of the development is limited. not everyone responds to treatment and of those people who do. for many established 157 . 2004a) there has been significant therapeutic innovation and research effort but in comparison with the research on pharmacological interventions. Recommendations are made where there is robust evidence to support the effectiveness of an intervention. 2006). are made in Chapter 8. Research recommendations that. People with depression typically prefer psychological and psychosocial treatments to medication (Prins et al. not everyone remains free of depression in the long term.. 2008) and value outcomes beyond symptom reduction that include positive mental health and a return to usual functioning (Zimmerman et al.

. The majority of available trials of psychological and psychosocial interventions have focused on the acute treatment of depression. these are normally guided by 158 . There is ongoing debate among researchers. 1995) set out in Figure 5.. 50 to 70% of patients in trials will achieve remission but a substantial proportion will go on to relapse (see Chapter 2).3 HOW DO PSYCHOLOGICAL AND PSYCHOSOCIAL INTERVENTIONS BECOME EVIDENCE BASED? For a therapy to become established as an effective treatment in routine care it typically passes through several phases of treatment development (Rounsaville et al. priorities for further research are suggested in order to establish more definitively which therapies work most effectively for people with depression. Where established therapies are not recommended. 2008). and of those who do.. 2006). 2008). 6. absence of evidence does not mean evidence of no effect. In the first phase of treatment development. Just because an approach is not recommended here does not mean that it is not effective or that it should never be provided. this does not necessarily mean that the withdrawal of provision from the NHS is endorsed but may suggest the need for further research to establish their effectiveness or otherwise. Several of the approaches considered below have shown consistently greater efficacy than control conditions in such trials. especially in supporting their longer-term recovery – a pressing concern for those people who experience recurrent depression. this should be contrasted with data from the STAR*D trial focusing on pharmacological treatments where remission rates in the initial phase of the study were between 28 and 33% (Trivedi et al. In the research recommendations (see Section 8. 2005). In other cases a weak or limited evidence base may lead to a qualified or restricted recommendation. The development of the evidence base is nicely illustrated by the ‘hourglass model’ (Salkovskis. usually of mild to moderate severity (although it should be noted that many of the participants in these trials will have had a number of previous episodes of depression). a substantial minority of patients do not respond adequately to treatment (both pharmacological and psychological). 2001. The likelihood of relapse will depend on the person’s history of depression and is higher in those with a significant past history of depression. Craig et al. with even the most effective treatments for depression. less than half of treated patients achieved full remission and sustained it over a period of 2 years following treatment (for example. therapists and policy makers about what constitutes the best evidence for psychological and psychosocial interventions and how this evidence should be used (Kazdin. However. Hollon et al. Typically. in one study of the psychological treatment of people with mostly chronic or recurrent depression (mean duration of episode: 46 months). However. a theoretical model and therapeutic approach are articulated. For example. However. a substantial proportion relapse.12).Introduction to psychological and psychosocial interventions therapies and promising new developments there will be insufficient evidence to recommend them. rather that the question of efficacy has not yet been adequately addressed to warrant a specific recommendation. As in most clinical sciences..

often in the form of a treatment manual42. For example. Manuals usually contain a mix of indicative as well as prescriptive elements. Often in this initial phase of treatment development. In many cases manuals also specify the frequency.Introduction to psychological and psychosocial interventions Figure 5: The hourglass model astute clinical observations and theoretical ideas about processes involved in the disorder. case reports. in cognitive therapy negative distortions in thinking were identified as key in maintaining depression. between treatment outcomes and outcomes for the group who did not receive treatment. in an unbiased manner. and. 159 . The neck of the hourglass represents the stage where a definitive RCT is conducted to establish efficacy. It sets out the theoretical rationale for the intervention and specifies the knowledge and skills required to deliver the intervention competently. and followed by interventions designed to target these processes. For example. Thus. and therapy therefore aims to help clients identify and respond to these distortions. the new treatment is compared with a meaningful comparison group. single case studies and expert opinion provide preliminary evidence that is used to refine the treatment approach. 1979). This exploratory trial lays the groundwork for a more definitive trial. intensity and duration of the intervention. since effective implementation of most interventions involves an element of clinical judgement. if ethically 42A treatment manual describes how an intervention should be delivered. clinical innovators develop novel treatment approaches. In healthcare research the RCT is considered the gold standard for establishing a treatment’s efficacy due to its ability to distinguish. help people become more active. and test out and change their cognitive distortions and underlying beliefs (Beck et al. an uncontrolled open trial enables preliminary research into the potential efficacy of a treatment.. If the treatment appears promising. Typically it contains an account of the disorder and/or problem to be treated and the specific population for which the intervention was developed. a treatment manual for cognitive therapy for depression sets out how to engage people. Through a process of careful experimentation and observation. this may include another active treatment.

with other comparators such as a placebo.. the first meta-analysis was conducted in 1990 (Robinson et al. The intention was to both inform the reviews undertaken for this guideline and provide a better understanding of the context for it. 1977. 6. or superior to. De Los Reyes & Kazdin... systematic literature reviews and meta-analyses can make sense of larger bodies of data and make inferences about which factors may moderate or mediate treatment effects.3. 2004. the GDG and technical team reviewed and evaluated not only relevant RCTs. the first RCTs were published in the late 1970s and early 1980s (Rush et al..1 Recent systematic reviews of psychosocial treatments for depression As part of the development of this guideline update. and is it acceptable to clients and therapists? Other research designs. while meta-analysis can be a powerful tool for synthesising the results of several studies. and the effectiveness and cost-effectiveness studies have only started to emerge in the last decade (Bower et al. over the past 50 years there has been a significant expansion of theories and therapies for depression.. However. 2005. These studies also drive the next incremental phases of clinical research. 2008.. another established treatment. only a relatively small number of these therapies have travelled the full empirical road and demonstrated that they are efficacious and can be cost-effective treatment options for the NHS. Rawlins.. Rush et al. Having established that the therapy is effective. Kazdin. 2003. As will be apparent from the summary below. 2008). the development work took place in the 1960s and 1970s. and routinely collected outcomes data.Introduction to psychological and psychosocial interventions justifiable. RCTs are explored and critiqued in detail elsewhere (Westen et al. is the therapy appropriate for routine care settings. To take the example of CBT. The most frequent challenges in interpreting meta-analyses 160 . an attentional control. It is beyond the scope of this chapter to discuss the RCT in detail and its role in evaluating psychosocial treatments. but also considered recent metaanalyses that had been published since the previous guideline. The final phase of treatment development is depicted in the bottom of the hourglass. the manual was published in 1979 (Beck et al. Stirman et al. 1981. This means that many therapies have not been subjected to a full test of their efficacy. Finally. may be suited to answering important questions at this stage. 2000. treatment delivery less adherent and treatment contexts more varied? In short. This enables the researchers to conclude that the new treatment is better than no active treatment and as good as.. In summary. do the outcomes hold up? Is the treatment acceptable and accessible? Can therapists be readily trained. 2008. The phases of treatment development illustrated in the hourglass demand considerable resource and time and this may explain the more limited evidence base for psychological and psychosocial interventions compared with pharmacological interventions. Byford et al.. 1979). as the evidence base accumulates.. it is not without problems. Kovacs et al. Scott et al. 1981). 2003). when the high internal validity expected in an RCT is traded for external validity. 1990). this phase of treatment development asks: is the treatment exportable to real world settings where therapist competence may be more variable. usual care or no treatment.

Cuijpers and colleagues (2008a) concluded that pharmacological and psychological interventions may be equivalent in major depression but that pharmacological interventions may be more effective for dysthymia than psychological interventions (see Chapter 13 for a review of interventions for subthreshold depressive symptoms). Ekers and colleagues (2007) reviewed the effectiveness of behavioural activation in the treatment of depression. short-term psychodynamic psychotherapy. 2004. seriously limited the ability of the data to support the conclusions drawn. One meta-analysis compared the efficacy of psychological and pharmacological interventions in the treatment of adult depressive disorders (Cuijpers et al. which would best inform a study of comparative effectiveness. dementia and postnatal depression). and the nature of the interventions compared (high. 2008. 2008b). This failure rests on two main issues: first.Introduction to psychological and psychosocial interventions are the nature of the studies selected for inclusion. interpersonal therapy (IPT).. Leichsenring and colleagues (2004). seven were considered of particular relevance and they are briefly summarised below. 2008). Finally. Abbass et al. 161 .. Leichsenring and Rabung (2008) and Abbass and colleagues (2008) concluded that psychodynamic psychotherapy is effective in the treatment of a broad range of mental health disorders (and by implication depression). the nature of the disorders reviewed (which included physical health problems.and low-intensity interventions were grouped together). This is largely supported by the available data. second. the trials they reviewed were designed to test differences in efficacy not establish equivalence (see Piaggio and colleagues [2006] for fuller discussion of this issue). Of the recent meta-analyses identified during the development of this guideline. Cuijpers and colleagues (2008b) concluded that there were no large differences in efficacy between psychological treatments for mild to moderate depression including CBT. a Cochrane review (Mead et al. this finding may reflect the fact that the dataset for pharmacological interventions is stronger (it has a more extensive set of high-quality studies and less heterogeneity) than that for psychological interventions. they had been considered for the reviews in this guideline. Leichsenring & Rabung. the approach to synthesising the data and the way the results are interpreted. A further meta-analysis looked at different types of psychological treatments and analysed their effectiveness in the treatment of depression (Cuijpers et al.. However. However. rather than it being due to a large number of high-quality head-to-head studies. behavioural activation43.. problem solving. All the meta-analyses were assessed for quality and the references from the included studies were checked to verify that. 2008a). where appropriate.. 2008) evaluated the effectiveness of physical exercise in the treatment of depression. Leichsenring and Rabung (2008) looked at long-term psychodynamic 43Note that these three interventions may be seen as belonging to a broad school of cognitive and behavioural therapies. social skills training and non-directive supportive therapy. a more accurate conclusion would be that Cuijpers and colleagues (2008a) had failed to find such differences rather than establishing that no differences existed. Three meta-analyses analysed the efficacy of psychodynamic psychotherapy in several mental health disorders including depression (Leichsenring et al.

mixed. 2004a) has been influential in reshaping the types of psychological and psychosocial treatments available for people with depression. High. Most notably there has been a recent increase in the accessibility of evidence-based therapies. in particular for people with common mental health disorders (Department of Health. again leading to caution in the way the results are interpreted.3. They state that because discontinuation from exercise can be substantial. which confounds interpretation of the data. it should be noted that these reviews contained very few studies of depression (three in total across the three reviews from which it was possible to extract data).and lowintensity behavioural activation. 6. and so on). 2007).and long-term psychodynamic versions of solution-focused psychotherapy outpatients with mood or anxiety disorders only 65. Ekers and colleagues’ (2007) review concluded that behavioural activation is an effective treatment for depression. they combined data on high. However. Alongside the NICE guideline and evidence 162 . Leichsenring and colleagues (2004) and Abbass and colleagues (2008) evaluated the effectiveness of short-term psychodynamic psychotherapy versus control groups ranging from medication management to psychotherapeutic support.and low-intensity behavioural activation. with outcomes superior to those of supportive counselling and brief psychotherapy. but noted that the effects are less convincing for those with an established diagnosis. In addition. For example in the Knekt and colleagues’ (2008) study of short. their conclusion should be treated with some caution for the following reasons: their analysis combined data from trials that included subthreshold symptoms.Introduction to psychological and psychosocial interventions psychotherapy compared with shorter forms of psychotherapy.2 Increasing the availability of psychological and psychosocial therapies in healthcare settings The previous guideline on depression (NICE. They reported a slow rate of recovery initially in the psychodynamic psychotherapy group and it is difficult to determine whether or not the long-term benefits associated with psychodynamic psychotherapy resulted specifically from the therapy or the prolonged contact with the therapist during that time. The fact that so few studies were concerned with depression significantly limits the validity of their conclusions in relation to this guideline. In Ekers and colleagues’ (2007) and a number of other reviews. They did not specify details about particular forms (that is. and the studies they included were not all peer-reviewed and did not meet the quality criteria established for this guideline. whether group or individual. Mead and colleagues (2008) concluded that physical activity should be recommended for people with depressive symptoms and those who fulfil the diagnostic criteria for depression. anaerobic. it is better to recommend a physical activity that the person will enjoy. are considered separately in this guideline. aerobic. there were a large number of patients in the study who had subsidiary treatments during the same period. or duration of exercise because of lack of consistent evidence. these interventions are combined in the meta-analyses. However. and other psychological interventions.8% had recurrent episodes of major depressive disorder.

First. 6. Another development from the previous guideline that formed part of the IAPT programme is the stepped care framework (see Chapter 5 for further details). Psychosocial and psychological interventions.3 Improving Access to Psychological Therapies initiative as an example of increasing the accessibility of established evidence-based therapies The IAPT programme (Department of Health. which became the organising principle for the provision of IAPT services. 11 further IAPT pathfinder sites began to explore the specific benefits of services to vulnerable groups. Many of those presenting to services will of course have chronic disorders and will. The IAPT programme began in 2006 with demonstration sites in Doncaster and Newham focused on therapies for adults of working age.and low-intensity psychological CBT-based interventions focused on new presentations to services and including the opportunity for self-referral.3. in the case of depression.Introduction to psychological and psychosocial interventions base. The development of IAPT was driven by an acknowledgement that the treatments NICE recommended were not as accessible as they should be and sought to redress this imbalance through a large investment of new training and service monies in the NHS. The initial focus of the programme is on high. The high direct and indirect costs associated with depression. there is greater understanding of how depression presents in the NHS and models of care and service delivery have been shaped accordingly (see Chapter 5). In 2009 it is expected that other interventions such as IPT will form part of the treatments offered by IAPT. Finally. A national rollout of IAPT delivery sites is now underway and is scheduled to complete in 2013. public demand and expectation influence service commissioners. A related element of the organisation of psychological therapies in the IAPT programme is the 163 . have also been drivers. a number of factors determine whether a psychological or psychosocial therapy becomes accessible in the NHS. and the suffering experienced by people with depression and their families and carers.000 patients per year. It is expected that it will lead to large increases in the accessibility of evidencebased psychosocial and psychological interventions. This has been one of the drivers for the development of less intensive therapies as well as innovative delivery formats such as group-based work. The NHS. In 2007. like all healthcare systems. User groups have long advocated the need for psychological and psychosocial approaches and this has influenced commissioning at a national and regional level. particularly high-intensity therapies that involve one-to-one therapy over longer periods of time. The goal is to alleviate depression and anxiety using NICE-recommended treatments and help people return to full social and occupational functioning. has a finite limit on its resources and there is therefore an impetus to find therapies that are as cost effective as possible. are resource intensive. require not just the treatment of the acute problems but also help with the prevention of relapse. 2007) supports Primary Care Trusts in England in implementing NICE guidelines for people with depression and anxiety disorders (similar programmes are underway in Scotland and Northern Ireland).600 therapists and treat a further 45. The intention is to provide £340 million of additional funding to train 3.

nature of depression.4. so too can some of the assumptions typically made in clinical practice (Kazdin. is poorly understood both for psychological interventions and pharmacological interventions (see Chapters 9 to 12). clinical guidelines are a guide for clinicians and not a substitute for clinical judgement. Lambert and Ogles (2004) and Roth and Pilling (2009). marital status. but some of the key factors that may influence treatment decisions are discussed below. In the same way that RCTs can be critiqued. including the impact of such tailoring on outcomes. expectations and preferences and experiences of previous treatment. which often involves tailoring the recommendation to the needs of the individual. 1991). This approach is consistent with that taken in all clinical guidelines and is set out in Chapter 1 of this guideline.. Interested readers can refer to several texts for a more detailed review. training. 6.and low-intensity is adopted in this guideline and is the basis on which Chapters 7 and 8 are organised. inevitably they make recommendations about the average patient.2 Therapist factors Several therapist factors that could potentially affect treatment have been considered.Introduction to psychological and psychosocial interventions distinction between high-intensity psychological interventions (that is. This distinction between high. There is an increasing research literature addressing factors that can affect treatment choices and outcomes but the research has as yet produced little that directly relates to the outcome of psychosocial and psychological treatments for depression. 6. including demographics. that is. including demographics. use of supervision and 164 . for example.4 CONTEXTUAL FACTORS THAT IMPACT ON CLINICAL PRACTICE Recommendations in this guideline are largely based on the syntheses of trial data from groups of patients with depression. In the main. few factors consistently predict treatment outcomes except chronicity and severity of depression. 6. Unfortunately the relationship of factors that may guide the tailoring of clinical practice recommendations to individual needs. formal psychological therapies such as CBT.1 Patient factors A broad array of patient factors that could potentially affect treatment choices have been considered. where a paraprofessional acts to facilitate or support the use of self-help materials and not to provide the therapy per se. IPT or couples therapy provided by a trained therapists) and low-intensity interventions such as CCBT. professional background. social factors and culture. Sotsky et al.4. which point to reduced treatment effectiveness across treatment modalities (for example. physical activity programmes and guided self-help. It is beyond the scope of this chapter to review these in depth. 2008).

although some therapists produce consistently 165 . Two related aspects are dealt with below. it is the consistency. a position reflected in many humanistic models where the therapeutic relationship itself is seen as integral to the change process. 1994) is often cited in support of this argument and is usually referred to as ‘the dodo-bird hypothesis’ (Luborsky et al.4 Therapist competence Studies of the relationship between therapist competence and outcome suggest that all therapists have variable outcomes. agreement about the goals or outcomes the therapy aims to achieve.4.Introduction to psychological and psychosocial interventions competence. 2008a). type of measures applied and the stage of therapy at which measures are applied).3 The therapeutic alliance There are various definitions of the therapeutic alliance. but clearly it is unlikely to be sufficient to account for the majority of the variance in outcome. However. characterised by a positive and mutually respectful stance in which both parties work on the joint enterprise of change. rather than the size of this correlation. 1991. since a correlation of 0. However. Rogers.. 2000). Horvath & Symonds. a finding that applies across a heterogeneous group of trials (in terms of variables such as type of therapy. 6. with technique relegated to a secondary role (for example.25 (for example. with some commentators arguing that technique is inappropriately privileged over the alliance. There has been considerable debate about the importance of the alliance as a factor in promoting change. and the quality of the bond between therapist and patient. Elkin.. Ahn & Wampold. which is most striking. The failure of some comparative trials to demonstrate differences in outcome between active psychological therapies (for example. Therefore it seems reasonable to debate the extent to which a good alliance is necessary for a positive outcome of an intervention. but in general terms it is viewed as a constructive relationship between therapist and client. 1975). patient presentation. 2006) but should also include a consideration of cost effectiveness as well as clinical efficacy (NICE. 6. for example Feeley and colleagues (1999) reported that alliance quality was related to early symptom change. It should also be noted that the alliance is itself affected by the process of treatment. namely the therapeutic alliance and therapist competence. Stiles et al. apart from the fact that dodo-bird findings may not be as ubiquitous as is sometimes claimed this does not logically imply that therapy technique is irrelevant to outcome.4. Martin et al. Meta-analytic reviews report consistent evidence of a positive association of the alliance with better outcomes with a correlation of around 0. 2001. Identifying and interpreting equivalence of benefit across therapies remains a live debate (for example. 1951).25 would suggest it could account for only 6% of the variance in the outcome. Bordin (1979) conceptualised the alliance as having three elements comprising the relationship between therapist and patient: agreement on the relevance of the tasks (or techniques) employed in therapy..

the overall CTS score accounted for 15% of the variance in outcome. the CBT items were separated into two subscales labelled ‘cognitive therapy – concrete’ and ‘cognitive therapy – abstract’. Some understanding of what may account for this association emerges from three studies by DeRubeis’s research group (DeRubeis & Feeley. However. Feeley et al. subscales of which contained items specific to CBT.. higher levels of competence were associated with greater improvements in depressive symptoms. and the highest possible levels of training and experience are desirable for those therapists treating patients with severe. 2009). All of the studies made use of the Collaborative Study Psychotherapy Rating Scale (CSPRS: Hollon et al. especially with Axis II pathology. 2009). In routine 166 . Concrete techniques can be thought of as pragmatic aspects of therapy (such as establishing the session agenda. enduring or complex presentations.. 2001) are associated with competence and are independently associated with better outcomes (Burns & Nolen-Hoeksema. 1999. degrees of formal training (Brosan et al. 1996) and personality disorders (Davidson et al. what is it that therapists do in order to achieve good outcomes? A number of studies are briefly reviewed here. DeRubeis and colleagues (2005) found that the most competent therapists had good outcomes even for patients with the most severe levels of depression. 2004). This section. 2007) and experience in that modality (James et al. draws on a more extensive review of the area by Roth and Pilling (2009).. that is. There is evidence that more competent therapists produce better outcomes (Barber et al. Both DeRubeis and Feeley (1990) and Feeley and colleagues (1999) found some evidence for a significant association between the use of ‘concrete’ CBT techniques and better outcomes. 1989). Brotman et al. Okiishi et al. In patients with neurotic disorders (Kingdon et al. Although the simple correlation of the CTS with outcome suggested that it contributed little to outcome variance.Introduction to psychological and psychosocial interventions better outcomes (for example. have been associated with poorer outcomes in therapies. Although competence in psychological therapies is hard to measure in routine practice. The benefits of high levels of competence over and above levels required for basic practice has been studied in most detail in the literature on CBT for depression. Ratings of competence were made on the Cognitive Therapy Scale (CTS). On the basis of factor analysis. specifically... a subset of items on the CTS accounted for most of this association.. All therapists should have levels of training and experience adequate to ensure a basic level of competence in the therapy they are practicing. which focuses mainly on CBT and depression. when controlling for pre-therapy depression scores.. setting homework tasks or helping clients identify and modify negative automatic thoughts). 2003). In an early study. 1990. but the detrimental impact of these factors is lessened for highly competent therapists.. Shaw and colleagues (1999) examined competence in the treatment of 36 patients treated by eight therapists offering CBT as part of the National Institute of Mental Health in England trial of depression (Elkin et al. 1996. 1988).. Kuyken and Tsivrikos (2009) found that therapists who are more competent have better patient outcomes regardless of the degree of patient comorbidity. 2006. adherence and the alliance. regression analyses indicated a more specific set of associations.. In general. high severity and comorbidity. 1992). Kuyken & Tsivrikos. A number of studies have also sought to examine more precisely therapist competence and its relation to outcomes.

although modest.22). Bryant and colleagues (1999) examined factors leading to homework compliance in 26 patients with depression receiving CBT from four therapists. In a completer sample (N 21) better outcomes were associated with overall competence on the CTS (r 0. Of these. therapists should receive regular supervision and monitoring of outcomes. Overall there was a significant. the number of trials on which this comparison is made is small and any conclusions must therefore be tentative. with ratings of competence made on the CTS. and between outcome and homework compliance (r 0. As in other studies.Introduction to psychological and psychosocial interventions practice in services providing psychological therapies for depression. Kazantzis and colleagues (2000) report a meta-analysis of 27 trials of cognitive and/or behavioural interventions that contained data relevant to the link between homework assignment. but ratings of their general therapeutic skills. Trepka and colleagues (2004) examined the impact of competence by analysing outcomes in Cahill and colleagues’ (2003) study. in the full sample this association was only found with the ‘specific CBT skills’ subscale of the CTS. greater compliance with homework was associated with better outcome. 167 . A small number of studies have explored whether compliance with homework is related to better outcomes. the remainder were seen for a range of other problems. These results remained even when analysis controlled for levels of the therapeutic alliance. compared with four of the 11 clients treated by the less competent therapists. Six clinical psychologists (with between 1 and 6 years’ post-qualification experience) treated 30 patients with depression using CBT. There was also some evidence that compliance was increased if therapists checked how the patient felt about the task being set and identified potential difficulties in carrying it out. The focus of the research on both the alliance and therapist competence has been on high-intensity interventions but it is reasonable to expect that they are potentially of equal importance in the effective delivery of low-intensity interventions. 11 reported on the effects of assigning homework in therapy and 16 on the impact of compliance.36). All forms of CBT place an emphasis on the role of homework because it provides a powerful opportunity for patients to test their expectations. compliance and outcome. The type of homework varied. it was not so much therapists’ CBT-specific skills (such as skilfully assigning homework or providing a rationale for homework) that were associated with compliance. although this was usually by therapist report. and particularly whether they explicitly reviewed the homework assigned in the previous session.47). association between outcome and assigning homework tasks (r 0. although rather fewer have examined the therapist behaviours associated with better patient ‘compliance’ with homework itself. Using a stringent measure of recovery (a BDI score no more than one SD from the non-distressed mean). In terms of therapist behaviours. Agreeing and monitoring homework is one of the set of ‘concrete’ CBT skills identified above. as did the way in which compliance was monitored. nine of the 10 completer patients treated by the more competent therapists recovered. In 19 trials patients were being treated for depression or anxiety. While Kazantzis and colleagues (2000) indicate that homework has greater impact for patients with depression than anxiety disorders.

when first author only is used). CINAHL Database inception to January 2008 July 2008 RCT People with a diagnosis of depression according to DSM. Study IDs in title case refer to studies included in the previous guideline and study IDs in capital letters refer to studies found and included in this guideline update.934 participants. a new systematic search was conducted looking at both published and unpublished RCTs. each study considered for review is referred to by a ‘study ID’ made up of first author and publication date (unless a study is in press or only submitted for publication. 168 . or depressive symptoms as indicated by depression scale score for subthreshold and other groups Behavioural activation Cognitive behavioural therapies CCBT Counselling Couples therapy Guided self-help IPT Problem solving Physical activity Rational emotive behaviour therapy Short-term psychodynamic psychotherapy Treatments 6. PsycINFO. EMBASE.5 DATABASES SEARCHED AND INCLUSION/EXCLUSION CRITERIA For the guideline update. All trials were published in 44Here and elsewhere in the guideline. ICD or similar criteria.6 STUDIES CONSIDERED44 A total of 139 trials relating to clinical evidence met the eligibility criteria set by the GDG. 6. The electronic databases searched for published trials are given in Table 18 (further information about the search strategy can be found in Appendix 8). References for studies from the previous guideline are in Appendix 18 and references for studies for the update are in Appendix 17b. providing data on 12.Introduction to psychological and psychosocial interventions Table 18: Databases searched and inclusion/exclusion criteria for clinical effectiveness of psychological treatments Electronic databases Date searched Update searches Study design Population MEDLINE.

contain a number of studies (such as BROWN2001. 95 studies found in the search for this guideline update were excluded from the analysis. MANBER2008 and PASSMORE2006) that do not appear in the full guideline.6.3. 45Appendix 17b.1 and Section 8.2. Further information about both included and excluded studies can be found in Appendix 17b45.Introduction to psychological and psychosocial interventions peer-reviewed journals between 1979 and 2009. Four studies included in the previous guideline were excluded from this guideline update (see Section 7. CONSTANTINO2008. but they did not warrant inclusion in the full guideline because they did not show any clear results. These studies were not excluded because they still met the criteria for the review.1). Section 8. 169 .3. In addition. and also Appendix 16b and 19b.

170 . 1990). over the internet (Christensen et al. Early studies suggested that patients find computer-based treatment acceptable and they manifest degrees of clinical recovery of similar magnitude to those who have face-to-face therapy (Selmi et al. The technology more recently available has led to the development of a more sophisticated range of computer-based or internet-based CBT programmes.1 COMPUTERISED COGNITIVE BEHAVIOURAL THERAPY Introduction The use of information technology to deliver psychological treatments has been explored.1 7. It can be used as the primary treatment intervention with minimal therapist involvement or as augmentation to a therapist-delivered programme where the introduction of CCBT supplements the work of the therapist. for example self-help delivered by telephone (Osgood-Hynes et al. 2004). Computerised cognitive behavioural therapy (CCBT) programmes engage the patient in a structured programme of care. the content of which is similar to and based on the same principles as treatment provided by a therapist following a standard CBT programme. Cognitive behavioural therapy (CBT) is currently the main psychological treatment approach that has been computerised.1. In the review for this guideline the focus is on CCBT as a primary intervention and not as a means of augmenting therapist delivered treatment.Low-intensity psychosocial interventions 7 LOW-INTENSITY PSYCHOSOCIAL INTERVENTIONS This chapter reviews the evidence for the effectiveness of a range of low-intensity interventions. 7.. Most of the programmes have been developed to treat a range of depressive and/or anxiety disorders. guided self-help and physical activity programmes in the treatment of depression. 1998).. including computerised cognitive behavioural therapy (CCBT). often explicitly as part of a stepped-care programme. 2002) and by computer (Proudfoot et al. The programmes vary in style. Direct staff input is usually limited to introducing the programme.. brief monitoring and being available for consultation. DVD or the internet. The review in this guideline supersedes and updates the aspects of the technology appraisal concerned with depression. 2006a). Definition CCBT is defined in this guideline update as a form of CBT.. which covers both depression and anxiety disorders. which is delivered using a computer either via a CD-ROM. These have been the subject of a technology appraisal (NICE. degree of complexity and content.

online psychoeducation. an information control. 47Severe 171 . which also includes details of excluded studies48. waitlist control and treatment as usual (TAU). WRIGHT2005 was excluded as the GDG did not consider the intervention provided to be the same as CCBT provided in the NHS (that is. One study. Table 19: Summary study characteristics of CCBT studies CCBT versus control No. depression. moderate depression: 41% CCBT and 56% treatment as usual. of those people with depression and subthreshold symptoms allocation across the differing severity levels of depression (mild. mild depression: 20% CCBT and 11% treatment as usual. group CBT. a discussion control. WRIGHT2005 focused on CCBT augmentation of a therapist-delivered intervention). but results should be interpreted with caution as randomisation to the study was not stratified by diagnosis. It is important to mention that this sub-analysis gives an indication of the effect in a depressed sample. In this study. 48For this review studies from the previous guideline were re-entered into the study database for the guideline update in Appendix 17b. Summary study characteristics of the included studies are presented in Table 19 with full details in Appendix 17b. depression: 39% were assigned to the CCBT group and only 33% were assigned to treatment as usual. PROUDFOOT2004A. the review team and GDG calculated that only 39% of the population met criteria for major depressive disorder (MDD). trials 7 RCTs (Total participants) (1412) Study IDs (1) ANDERSSON2005A (2) CHRISTENSEN2004A* (3) CLARKE2002 (4) CLARKE2005* (5) PROUDFOOT2004A (6) Selmi1990 (7) SPEK2007* (1) 117/74/36 (2) 347/71/36 CCBT versus active comparator 2 RCTs (548) (1) CHRISTENSEN2004A* (2) SPEK2007* N/% female/mean age (years) (1) 368/71/36 (2) 191/63/55 Continued 46Study IDs in capital letters refer to studies found and included in this guideline update. Furthermore. subthreshold depression or anxiety.2 Studies considered46 Seven RCTs were included. The evidence is presented here for the full sample and a sub-analysis was also conducted including only those who met diagnostic criteria for depression.676 participants.1. Data were available to compare CCBT with traditional CBT. moderate and severe) was not balanced47. included a population of people with mixed depression and anxiety. providing data on 1.Low-intensity psychosocial interventions 7.

4) CCBT versus active comparator (1) MoodGYM (2) Coping with Depression CCBT Support (1) Phone to direct website use by lay interviewer Continued 172 . 12 KPDS (CES-D 21. 11% minor depression. 12 KPDS (21.4) CCBT programme (1) Not fully described.Low-intensity psychosocial interventions Table 19: (Continued) CCBT versus control (3) 223/75/44 (4) 200/77/47 (5) 274/74/44 (6) 36/64/28 (7) 201/63/55 Diagnosis (average (1) Major depression (20.5) (5) Depression and/or anxiety disorders (25. 12 EDS (18.0. 19% intermittent depressive disorder (22.0) (6) 69% major depression.0) (4) Depression (22% no formal diagnosis) (CES-D 30. depression-only group 30. only mention is that it is based on Beck’s cognitive therapy (2) MoodGYM (3)–(4) Overcoming Depression on the Internet (5) Beating the Blues (6) Not reported (7) Coping with Depression (1) Email feedback from therapist (1) No formal diagnosis.5) (2) No formal diagnosis.5) (7) No formal diagnosis.5) (3) Depression (25% no formal diagnosis) (CES-D 31.7) baseline BDI) (2) No formal diagnosis. 12 EDS (18.

3 Clinical evidence Evidence from the important outcomes and overall quality of evidence are presented in Table 20 and Table 21. 3. respectively. The full evidence profiles and associated forest plots can be found in Appendix 16b and Appendix 19b.1.Low-intensity psychosocial interventions Table 19: (Continued) CCBT versus control CCBT versus active comparator (2) Phone to direct website (2) No support use by lay interviewer (3) Email reminders (4) Phone/postcard reminders (5) Nurse facilitating use at clinic (6) Support available at start and end of sessions (7) No support Comparator (1) Online discussion group (2) Weekly phone discussion (3)–(4) Health information webpage (5) TAU (6)–(7) Waitlist (1) 10 weeks (2) 6 weeks (3) Mean 32 weeks (4) Mean 16 weeks (5) 8 weeks (6) 6 weeks (7) Not reported (1) 6 months (2) 6 and 12 months (3)–(4) Not reported (5) 2. 8 months (6) 2 months (7) 12 months (1) BluePages psychoeducation website (2) Group CBT: Coping with Depression course Length of treatment (1) 6 weeks (2) 10 weeks (mean) Follow-up (1) 6 and 12 months (2) 12 months *3-armed trial. 5. 7. 173 .

23 (CI 1.61 (CI –1.62 (CI –0. K participants 1.91 to –0.23 –0.22 to 0) (CI –0.33) At 3 months: SMD –0.93) Low 202 CCBT 01.27 (CI –0.28) CCBT versus discussion control CCBT versus information control Not reported CCBT versus any control Not reported – – – At endpoint SMD SMD –0.22) CCBT 01. n Forest plot number CCBT 01.01) .40 (CI –0. n High 477 RR 2.01 SMD –0.16) Quality Low Number of studies.54 to 0.58 to –0.82 (CI 0.40 (CI –0.57 to 1.20 (CI –0.174 Table 20: Summary evidence profile for CCBT versus control CCBT versus TAU control RR 1.43 to –0.11) K 1.01 Low-intensity psychosocial interventions CCBT versus waitlist control Leaving study early for any reason RR 0.01 Depression selfreport measures at endpoint SMD –0.35 (CI 0.95 to 1.70 to –0. n 274 K 2.46 to 0.06) – – – SMD –0.51 to 3.02) At 6 months: SMD –0.

73 to –0.56 (CI –0.03 1.02 CCBT 01. n 1146 Low-intensity psychosocial interventions Forest plot number CCBT 01. n 380 237 420 K 2.23 (CI –0.03 CCBT 01.02 CCBT 01.11) At 8 months: SMD –0.At 5 months: SMD –0. n 1. n 1. n K 7. n 369 Low Low High High High At 12 months: SMD –0.03 CCBT 01.04) Quality Low Number of studies. n 1. n 2.02 175 .02 CCBT 01.02 CCBT 01.03 CCBT 01.27) Low Moderate Moderate Moderate 202 K K K K CCBT 01.43 to –0.03 CCBT 01. n 195 178 164 186 K K K 2.42 (CI –0. K participants 1. n 1.85 to –0.

22 to 0.12) (25% versus 15%) (34% versus 43%) Moderate K 1. n 201 CCBT 05.4 Clinical evidence summary Seven studies included a comparison of CCBT with control groups. Two further studies also compared CCBT with an active comparator: CHRISTENSEN2004A compared CCBT with a psychoeducation website and SPEK2007 compared CCBT with group CBT (delivered by a therapist).02 SMD –0. participants Forest plot number SMD 0.31) Low K 1.02 CCBT 06.05 (CI –0.1. information control and discussion control. n 402 CCBT 06. The patients in the trials included in this review were drawn predominantly from groups in the mild-to-moderate range of depressive symptoms (mean baseline BDI scores between 18 and 25). Approximately half (53%) met diagnostic criteria while the remainder had no formal diagnosis.56 to 1.06 (CI –0. n 221 K 2.02 (CI –0.17) Low Depression self-report measures at follow-up Quality Number of studies. participants Forest plot number Depression self-report measures at endpoint Quality Number of studies.02 SMD 0. n 347 Low K 1. The control groups were varied: waitlist control.34) Low K 1.79 (CI 0.22 to 0.Low-intensity psychosocial interventions Table 21: Summary evidence profile for CCBT versus active comparator CCBT versus psychoeducation control Leaving study early for any reason Quality Number of studies.21 to 0.67 (CI 1.27 to 0.08 to 2. the results for depression scores at endpoint indicated a significant small-to-medium 176 . n 201 CCBT 04.02 6 months’ follow-up CCBT 05.59) RR 0. When studies including a non-active control group were analysed together.02 12 months’ follow-up SMD –0. n 276 CCBT 04.2) Low K 1.02 7. treatment as usual.03 (CI –0. participants Forest plot number CCBT versus group CBT control RR 1.

58. 95% CI –0. 0. 2008). and depression only.1..27) at 8 months’ follow-up. 7. physiotherapists and dieticians). reported the results of CCBT in a population with depression and anxiety. at 3 months: SMD 0. anxiety only. contacts with other services (chiropodists. The cost of buying the license to use Beating the Blues (plus 177 . 0. 95% CI –0. 0. this result was not clinically important. 2003. district nurses and health visitors).40.02. and at 5 months: SMD 0. 95% CI –0.40. –0. the evidence was more limited.21. –0. day surgery and accident and emergency attendance).Low-intensity psychosocial interventions effect size (SMD –0.22. anxiolytics and sedatives).45..70.99). Kaltenthaler et al. CCBT had a very small effect in reducing depression self-report scores (SMD –0. Also. the sub-analysis revealed that there was insufficient evidence to determine a significant effect of CCBT (at endpoint: –0.73. practice nurses. In terms of the effectiveness of CCBT at follow-up. 95% CI –0. when CCBT was compared with active controls (psychoeducation and group CBT) and results were observed at endpoint. Details on the methods used for the systematic search of the economic literature are described in Chapter 3.35. these were estimated using the method described by Lave and colleagues (1998). The results indicate that CCBT had a significant small/medium-sized effect (SMD –0.65. contacts with primary care staff (GPs. 95% CI –0. however.85.90. mixed depression and anxiety or anxiety disorders treated in the UK primary care setting. However. The paper by McCrone and colleagues (2003) compared the Beating the Blues software package versus standard care in people with a diagnosis of depression. depression-free days and quality-adjusted life-years (QALYs). –0. Costing was conducted prospectively on a subsample of the patients included in the RCT.5 Health economic evidence and considerations The systematic search of economic literature undertaken for the guideline update identified two studies on CCBT for people with depression set in the UK (McCrone et al. The study adopted a societal perspective. 95% CI –0. community mental health nurses. Section 3. when the mixed depression and anxiety and anxiety only populations were removed.11) at 5 months’ follow-up. The benefit measures used in the economic analysis were improvements in BDI scores.6. One study. contacts with home helps.22).11) in reducing self-reported depression scores when compared with treatment as usual at 3 months’ follow-up. 0. and medication (antidepressants.39. 95% CI –0.17). contacts with hospital services (inpatient care for psychiatric and physical health reasons.1. PROUDFOOT2004. Costs included: contacts with mental healthcare staff (psychiatrists.10.42. a significant small/mediumsized effect (SMD –0.19. counsellors and other therapists). and a significant medium-sized effect (SMD –0. –0. 95% CI –0. Evidence from two studies (CHRISTENSEN2004A and SPEK2007) that compared CCBT with an active control showed that at 12 months’ follow-up. psychologists.56. no clinically important difference was identified. outpatient care. The study was based on an RCT (PROUDFOOT2004A). favouring CCBT in patients with a range of severity of depressive symptoms. It should be pointed out that this study was of a population of mixed depression and anxiety.

This cost was determined using the throughput levels that were based on assumptions about the number of patients likely to be picked up from a general practice. Patients who complied with treatment were then assumed to be distributed across the four depression severity categories depending on the success of the intervention: minimal. Overcoming Depression and Cope. The price of the computer programme licence was obtained from the manufacturer. the utility value was selected from a study that combined the values from a number of different published studies using a range of sources and methods.000 per QALY. mild to moderate. the societal perspective was adopted. The following limitations of the study were noted: sensitivity analysis was conducted only on the cost of the CCBT programme. CCBT was one of the depression products and this was compared with care usually received in primary care. then there would be an 80% chance of the intervention being cost effective. A proportion of these were assumed to complete the treatment. These throughput levels were highly uncertain because of the novel nature of CCBT in the NHS. the probability of the intervention being cost effective was 85%. which is not recommended by NICE (2009a). which 178 . At a willingness-to-pay of £5. The same model structure was used to evaluate all three depression programmes. Variation in cost effectiveness by severity of depression was also explored with a subgroup analysis. Results were presented in the form of cost-effectiveness acceptability curves (CEACs). and the time horizon spanned just 8 months. Patients were given either CCBT or standard care over a 2-month period. The study may benefit from more scrutiny of this uncertainty by the use of sensitivity analysis. In terms of depression-free days. as this was deemed to be the most uncertain factor. which may have led to an underestimation of the potential costs and benefits of the intervention. At a cost effectiveness of £15. The economic analysis for the health technology appraisal by Kaltenthaler and colleagues (2008) aimed to evaluate a range of CCBT packages for the treatment of depression. which demonstrate the probability of an intervention being cost effective at different levels of willingness-to-pay per unit of effectiveness (that is. The study population consisted of patients with mild to moderate. In addition. moderate to severe or severe depression. The CEAC showed that the probability of Beating the Blues being cost effective over standard care was greater than 80% at a value of £40 per unit reduction in BDI score. The authors concluded that Beating the Blues had a high probability of being cost effective. The time horizon of the analysis was 8 months. the CEAC suggested that if society placed a value of £5 on a depression-free day. moderate to severe and severe. The indirect method in which QALYs were estimated was also problematic.000 per QALY. At the end of the 6-month period.Low-intensity psychosocial interventions overheads) was also considered. The software packages considered included Beating the Blues. the probability of Beating the Blues being cost effective was found to be 99%. at different cost-effectiveness thresholds the decision-maker may set). These packages were compared with treatment as usual in primary care over an 18-month time horizon. Patients were assumed to spend 6 months in their new severity state following treatment. The decision tree model compared two arms (CCBT and standard care). Those who did not complete CCBT were assumed to be offered standard care and this resulted in a set of transition probabilities between disease severity categories. in particular.

clinical support. patients who improved stayed the same or relapsed. depending on the level of severity of depression.6% for Cope and 54. Utility values were obtained from a dataset from a recently published UKbased RCT of supervised self-help CBT in primary care by Richards and colleagues (2003). This was based on the following assumptions: there are 10. with all products offering a fixed fee for purchase at the level of general practice. CORE is a self-report questionnaire of psychological distress that has been mapped onto the BDI. 2000). for Cope it was £7. patients were assumed to move between severity categories as stated previously over the next 2 months and then stabilised for the remaining 6 months of the model. capital overheads and training of support staff. The estimated costs of each intervention included licence fees. All three packages for depression demonstrated an incremental cost-effectiveness ratio (ICER) well below the cost-effectiveness threshold of £20.4% for Overcoming Depression. screening of patients. However.391. If they did not improve in the first place (they were in moderate or severe categories). for Beating the Blues it was estimated that 100 patients would come forward each year in practices of one to five GPs. There is considerable uncertainty surrounding these assumptions and this is one of the main drivers of cost. The mapping function was fitted to these data to provide BDI data on each case.000 patients per practice. In the second cycle.. the primary outcome in the CCBT studies. Beating the Blues was the only product based on an RCT. computer hardware. If they relapsed. The authors aimed to find utility values for depression linked to the BDI. then at 10 months after initiating treatment they were offered either another course of CCBT or treatment as usual in the CCBT arm. The incremental cost per QALY of Beating the Blues over treatment as usual was £1. Based on the estimated BDI scores. so the cost per patient depends on the number of patients likely to use each copy. The authors made assumptions about the throughput levels used to estimate the cost per patient using the programme and about the number of patients likely to be picked up from a general practice. 179 . they also stayed in the same severity category. and 10% of these will be treated each year. This study incorporated the EQ-5D and CORE (Evans et al. mild (BDI score 10 to 18).000 was 86. The probability of accepting Beating the Blues over treatment as usual at £30.801. moderate (BDI score 19 to 29) and severe (BDI score 30 to 63) depression and then linked each category with an average EQ-5D score based on people’s responses in each category.139 and for Overcoming Depression it was £5. Effectiveness estimates in terms of transition probabilities were sourced from published and unpublished trials for each of the products and further assumptions. Beating the Blues was found to be more effective and more costly than treatment as usual. The ranges of scores were reported to be comparable to those found in other studies. The study adopted the perspective of the health service. The subgroup analysis found no differences across the severity groupings. Each product has a licence fee tariff.000/QALY. Costs included intervention costs as well as other service costs. If they did not relapse they stayed in the post-retreatment severity category. 62. The licence fee is fixed.Low-intensity psychosocial interventions was 8 months after treatment began.8%. 1000 of these have depression. For example. Kaltenthaler and colleagues (2008) categorised patients in this dataset as having minimal (BDI score of 9).

primary data using generic preference-based measures in the relevant population would have been ideal. Summary of health economic evidence Beating the Blues was found to be more cost effective than standard care. comparison of the effect sizes in each case indicates that the various CCBT packages may offer similar benefits to people with depression compared with a baseline treatment such as waitlist control and treatment as usual. the economic analysis of Beating the Blues cannot be updated. Moreover. The results are based on indirect evidence as no head-to-head trials were identified. a number of additional trials of CCBT were found and the clinical-effectiveness data reviewed from these trials suggest that other CCBT packages (both CD-ROM and web-based) may be similarly effective to Beating the Blues. it was also the package that demonstrated the highest probability of being cost effective at £30. In addition. There remains scant evidence on the likely uptake in practice. for Beating the Blues. according to the economic analysis for 180 . A major cost component of Beating the Blues was its licence fee. the clinical trials used different comparators and outcome measures. although the model assumed more realistic throughput levels. QALYs were estimated from a population of patients receiving CBT. Based on the clinical and cost-effectiveness findings of Kaltenthaler and colleagues (2008). Since the publication of the technology appraisal on CCBT. Beating the Blues was recommended by NICE (2006a) as suitable treatment for patients with depression. One of the limitations of the economic model was that a number of parameters such as compliance and relapse rates were based on assumptions because of lack of relevant data. The problem of paucity of data mentioned in Kaltenthaler and colleagues (2008) remains and. Subsequently. Although no further cost-effectiveness analyses were identified. other CCBT packages considered in the clinical review are likely to incur lower intervention costs compared with Beating the Blues. Nevertheless. Regarding costs. relapse rates or costings have been made available since then. Moreover.000/QALY. However. The authors highlighted this as a strong assumption that needs validation and requires some caution when reviewing the findings of the Kaltenthaler and colleagues’ (2008) report. which suggests caution in making any inferences regarding the relative effectiveness of CCBT packages. therapist-led CBT relapse rates were used as an estimate for CCBT relapse rates.Low-intensity psychosocial interventions Beating the Blues was the sole package to be evaluated in the context of an RCT with a control group. Beating the Blues was the only package recommended in the technology appraisal. For example. Therefore. there remains a large amount of uncertainty regarding the costs of the licence per patient. an analysis of a depression only sample (some of the participants in the PROUDFOOT2004A trial had a diagnosis of anxiety) undertaken as part of this review suggests further caution in interpreting the outcomes of the trial. This study was based in the UK and therefore would be representative of those patients utilising the NHS. no new Beating the Blues RCT data has become available and there have been no new published economic evaluations in the UK related to Beating the Blues or other CCBT packages. This is due to uncertainty regarding the throughput of people receiving CCBT. no data on compliance.

and it is not within the scope of this guideline to make recommendations on specific self-help manuals. the RCTs of some web-based programmes describe minimal supervision requirements. it is possible that other service costs associated with provision of CCBT are similar across the packages. the costs of providing this intervention are going to be further reduced.to 10-minute telephonic contacts from lay interviewers to participants to assist in the use of the site. and could support the more effective targeting of resources.1 GUIDED SELF-HELP Introduction Guided self-help is generally accepted as being more than simply giving patients literature to read (this simpler alternative is usually referred to as pure self-help). In addition to intervention costs. as suggested by the findings of the clinical review. Moreover. People may prefer to use CCBT in the privacy of their homes. at home or at a public library). 7. It is potentially cost effective for patients with milder disorders. Service user preference is important. In the US there are over 2000 self-help manuals of different sorts published each year. overheads and supervision. some may prefer visiting their GP practice to access CCBT. See Richardson and colleagues (2008) for a review of publicly available guided self-help materials in the UK. as they do not include hardware and overhead costs. however. Furthermore. Contact with professionals is limited and tends to be of a supportive or facilitative nature. this may support a greater range of service user choice. The technology appraisal has shown Beating the Blues to be more cost effective than treatment as usual. then they could be also more cost effective than usual care. On the other hand. 2006a]). However. 181 . using conservative estimates of the likely uptake of the intervention. If other CCBT packages are similarly as effective as Beating the Blues (as indicated in the clinical review) and incur lower intervention costs. where patients can access a CCBT programme over the internet or at locations other than at a GP practice (for example. there is little published evidence on this topic regarding CCBT. Hardware and overheads are fixed costs and would be the same for both free and licensed programmes. the licence fee for Beating the Blues comprised 73% of the total intervention cost (see Appendix 15 of this guideline update and page 159 of the technology appraisal [NICE. free packages such as MoodGYM do not require a licence fee and therefore intervention costs are greatly reduced. providing this service would incur costs for hardware. If a web-based programme were to be offered at a GP practice.Low-intensity psychosocial interventions the technology appraisal. Most of the early literature on guided self-help came from the US. and often is based on a cognitive or behavioural psychological approach.2 7. and others with mobility problems may value the flexibility it offers. for example MoodGYM trialled by Christensen and colleagues (2004) described 6.2. other costs associated with the care of people with depression need to be assessed. if different packages result in similar improvements for people with depression. but rather the principle and practice of guided self-help in the NHS and related services. By offering a range of options for accessing CCBT.

usually to no less than three contacts and no more than six. which makes use of a range of books or other self-help manuals derived from an evidence-based intervention and designed specifically for the purpose. Two of the studies included in the original review (Bowman1995.2 Studies considered49 Sixteen studies were identified and included in the review of guided self-help. graduate and low-intensity workers in mental health) facilitates the use of this material by introducing. guided self-help is defined as a self-administered intervention designed to treat depression. dropouts were replaced. breastfeeding or personal preference. Definition For the purposes of the guideline.2. guided self-help can be a more accessible and acceptable form of therapy. 2004a). The included studies were grouped based on the nature of support offered to patients. This intervention would have no other therapeutic goal and would be limited in nature. Gellatly and colleagues (2007) considered guided self-help to include no more than 3 hours of input from a coach or guide. such as a requirement of a certain reading ability and understanding of the language used. drug interactions. particularly with written materials. therefore the review was substantially revised. many patients are not keen on using medication. and 44% will not read a book in any year (NCES. pregnancy. only nine of these had been identified and included in the previous guideline (NICE. A healthcare professional (or paraprofessional. On the other hand. For example. Carers and family members can also be involved in understanding the nature and course of depression through the material made available. The majority of guided self-help programmes are in book form and this review is limited to studies of such programmes.Low-intensity psychosocial interventions Guided self-help has some obvious limitations. 7. For those people. because of antidepressant intolerance. and many patients are understandably worried about having a formal diagnosis of depression recorded in their medical records. 22% of the US population is functionally illiterate. The main reasons for exclusion were not being an RCT and participants not meeting diagnostic criteria. for example. Data were available to examine the following strategies compared with waitlist or treatment as usual: ● individual guided self-help – with frequent therapist/coach support (10 to 50 minutes per session) – with frequent but minimum duration support (not more than 2 hours overall) 49Study IDs in title case refer to studies included in the previous guideline and study IDs in capital letters refer to studies found and included in this guideline update. Sixteen of the new studies found were also excluded. 182 . References for studies from the previous guideline are in Appendix 18. Wollersheim1991) were excluded in the revised review for this guideline because they no longer met inclusion criteria: in the Bowman1995 study. monitoring and reviewing the outcome of such treatment. 1997). and the Wollersheim1991 study had less than ten participants in each condition.

42. There is insufficient evidence to indicate a clear effect of either group. approximately 50% of the included population was aged 15.4 Clinical evidence summary Overall. Section 3. 0.77.2. 95% CI –0. 7. reports similar results when comparing individual guided self-help with support of frequent but minimum duration when compared with treatment as usual at endpoint (SMD –0.12).27. –0.88.2.98. Only two studies compared individual guided self-help with frequent and long-duration tutoring with control groups (Brown1984 includes a waitlist control comparison and LOVELL2008 includes a treatment as usual comparison). when compared with treatment as usual (BDI scores: SMD –0. the results are not significant and the CIs are wide when compared with waitlist control (BDI scores: SMD –0.Low-intensity psychosocial interventions ● ● group guided self-help/psychoeducation self-help with support by mail. –0. It is important to note that in one of the studies. 183 .49.21) and at 12 months’ follow-up (SMD –0. WILLIAMS2008. One study. –0.34). –0. Details on the methods used for the systematic search of the economic literature are described in Chapter 3. Then at 6 months. Summary study characteristics of the included studies are in Table 22 with full details in Appendix 17b.70.50.47). The full evidence profiles and associated forest plots can be found in Appendix 16b and Appendix 19b.08. with wide CIs.1.57. 7. Three studies looked at the effectiveness of self-help with support by mail only. However.2. the evidence indicates that guided-self help has a beneficial effect in people with both mild depression and subthreshold depression.02. but the data is insufficient and the CIs are too wide to reach any clear conclusions.14).62. 0.28.32. 95% CI –0. 95% CI –1. STICE2007. 95% CI –0. two studies indicate a small effect (SMD –0. 95% CI –0. The results at shorter follow-up periods (1.6. While the effect favoured individual guided self-help with support.53) and similarly. 7.and 3month follow-up) were not significant. there is clear evidence from five studies to indicate that individual guided self-help with support of frequent but minimum duration has a large effect in reducing depressive self-reported symptoms when compared with waitlist control (SMD –0.5 Health economic evidence and considerations No evidence on the cost effectiveness of individual or group-based guided self-help programmes for people with subthreshold or mild to moderate depression was identified by the systematic search of the health economics literature.3 Clinical evidence for guided self-help Evidence from the important outcomes and overall quality of evidence are presented in Table 23. which also includes details of excluded studies.02). respectively. –0. 95% CI –1. One medium-sized study reports BDI scores at endpoint indicating a medium effect (SMD –0. 95% CI –1. Two studies included group guided self-help.

Table 22: Summary study characteristics of studies of guided self-help Individual guided self-help (minimal support) 10 RCTs (904) (1) Beutler1991 (2) Brown1984* (3) FLOYD2004 (4) Jamison1995 (5) Landreville1997 (6) Schmidt1983* (7) Scogin1987 (8) Scogin1989 (9) WILLEMSE2004 (10) WILLIAMS2008 (1) 63/70/47 (2) 30/55/37 (3) 46/76/68 (4) 80/84/40 (5) 23/87/40 (6) 34/86/42 (7) 29/79/70 (8) 67/85/68 (9) 216/66/42 (10) 281/68/41 (1) 120/93/NA (2) 319/73/44 (3) 32/86/42 (1) BROWN2004 (2) HANSSON2008‡ (3) Schmidt1983* 3 RCTs (495) 3 RCTs (368) (1) GEISNER2006 (2) SALKOVSKIS2006 (3) STICE2007† Group guided self-help (psychoeducation) Self-help (with support by mail) 184 (1) 177/70/19 (2) 96/80/40 (3) 95/70/18 Individual guided self-help (with support) No. of participants) 2 RCTs (89) Low-intensity psychosocial interventions Study IDs (1) Brown1984* (2) LOVELL2008 N/% female/mean age (1) 30/55/37 (2) 59/73/38 . trials (no.

BDI 10 (3) No formal diagnosis. ES-D 20 Low-intensity psychosocial interventions Intervention (1) Coping with Depression (CWD) with individual support (2) Individual guided-self help 185 Continued . HAMD 10 (9) Subthreshold depression (10) No formal diagnosis. BDI 14 (1) Self-directed therapy (2) CWD (3)–(5) Bibliotherapy (Feeling Good) (6) Bibliotherapy (self-help manual) (7)–(8) Bibliotherapy (Feeling Good) (1) Psychoeducation workshop (2) Psychoeducation Contactus (3) Self-help group (large) (1) Personalised feedback and brochure with coping strategies by mail (2) Tailored workbook (3) Bibliotherapy (Feeling Good) (1) No formal diagnosis. (1) No formal BDI 10 (70%) diagnosis.Inclusion criteria/ diagnosis (1) MDD (2) GP diagnosis BDI 14 (1) MDD (2) MDD/subthreshold depressive symptoms/ intermittent depressive disorder (3)–(4) MDD (5) 74% MDD/26% minor depressive disorder (6) No formal diagnosis. BDI (2) Depression 14 (3) BDI 10 (2) No formal diagnosis. BDI 10 (7)–(8) No formal diagnosis.

186 Table 22: (Continued) Individual guided self-help (minimal support) (9) Minimum contact therapy (based on CWD course) (10) Guided self-help (1) Group CBT/focused expressive psychotherapy (2)–(8) Waitlist (9)–(10) TAU (1) 20 weeks (2) 8 weeks (3) 4–12 weeks (1) Waitlist (2) TAU (3) Waitlist (1)–(3) Waitlist Group guided self-help (psychoeducation) Self-help (with support by mail) (1) 1 day (mean) (2) 6 weeks (mean) (3) 8 weeks (1)–(2) Not reported (3) 30 days (mean) Low-intensity psychosocial interventions Individual guided self-help (with support) Control (1) Waitlist (2) TAU Length of treatment (1) 8 weeks (2) 12 weeks (mean) .

†3-armed trial. 187 . 12 weeks and 6 months (3) 6 months Follow-up (1) 1 and 6 months (2) Not reported Low-intensity psychosocial interventions *4-armed trial. ‡Cluster randomised trial analysed separately.(4)–(5) 4 weeks (6) 8 weeks (7) 4 weeks (8) 1 month (9) 60 days (mean) (10) 120 days (mean) (1) 3 months (2) 1 and 6 months (3)–(4) 3 months (5) 6 months (6) 10 weeks (7) 1 month (8) 6 months (9)–(10) 12 months (1) 3 months (2) Not reported (3) 10 weeks (1) 4 weeks (2) 4 weeks.

01 Depression self-report measures at endpoint SMD –0.01 1.Table 23: Summary evidence profile of guided self-help Individual guided self-help (minimal support) versus waitlist versus TAU versus waitlist versus TAU versus waitlist Group guided self-help (psychoeducation) Self-help (with support by mail) 188 versus TAU RR 7.02 to –0.34) (0.03) Individual guided self-help (with support) versus waitlist Leaving RR 0. 21 K n Moderate Moderate Moderate Moderate 1.77 (0.71 (0.95 to 55.01 RR 10.88 to 0.62) (1.28 SMD –0.56) Moderate K 3.70 to –0.62 to 4.14) .49 (–1.47) (–0.50 study early (0. n 368 GSH 07.08 (–0.01 SMD –0.01 GSH 03.24 RR 1.01 GSH 01.69) Low K n GSH 02.07) SMD –0.12) At 1 month: SMD –0.05 to 0.0 Not estimable RR 2. 227 K n 2.62 to –0.02) SMD –0.16 RR 1.50 to –0.51 (–1.05 to 4.45 (–1.21) (–0.32 (–0. 59 K n 6.42) At 6 months: SMD –0. 319 GSH 06.30 to 0.26) (0.98 SMD –0.75 to 31281.08 to 4.02 (–0. 497 K n 1.21) At 12 months: SMD –0.67 to 4.01 GSH 05.53) (–0.13) At 3 months: SMD 0. participants K n 1.56 to 0.27 SMD –0.94) for any reason Quality Low Low-intensity psychosocial interventions Number of studies.34) (–1.67 SMD –0.83 to At 3 months: –0.38 to 0.57 (–1. 30 Forest plot number GSH 04.08 to 0.77 to –0.42 (–0.

02 GSH 03.03 GSH 05.02 GSH 01.90 to –1.05 95 358 96 191 Number of studies. participants – – Low-intensity psychosocial interventions Forest plot number – – – – 189 . 204 1. 42 K n 5.03 GSH 07. n GSH 07.04 GSH 06.18) High K n GSH 01.Quality Low Low Moderate Moderate Moderate Low Low Moderate Moderate Moderate Low High K K K K 1.04 – 4.02 Depression Not reported clinician-report measures at endpoint Not reported SMD –1.02 GSH 03. participants GSH 02. 122 Forest plot number GSH 04. 204 K n K n 1.02 GSH 07. 161 – – – – Not reported – – Not reported Not reported Not reported Quality – – – – – Number of studies. n 1. n 3. 24 K n 1. 55 K n 1. 159 K n K n 1. n 2.02 SSH 05. 21 1.54 (–1.02 K n 1.04 GSH 07.

The clinical evidence suggests that individual guided self-help is effective in reducing self-reported depression scores when compared with waitlist control or treatment as usual. In addition. The unit cost of a low-intensity therapy worker is not currently available but was estimated to be comparable to that of a community mental health nurse at AfC Band 5. and so this was used to estimate total staff costs. which currently costs £4. salary on-costs. it is difficult to assess how these clinical improvements can be translated into overall improvements in patient health-related quality of life (HRQoL) that can used in a cost-effectiveness analysis. overheads and capital overheads plus any qualification costs. However. except that each patient would receive an individual copy of the self-help manual. Based on the assumption of there being five to six people per group. as part of their treatment each person receives a copy of the self-help manual. The cost-effectiveness of individual self-help also depends on the impact on downstream resource use and not just the service costs of delivering the interventions.3 7. 190 . There is a growing body of literature primarily from the US examining the effects of physical activity in the treatment of depression. The unit cost of an AfC Band 5 community mental health nurse is £51 per hour of patient contact in 2007/08 prices (Curtis. The intervention could be delivered by a mental health professional or a paraprofessional with each session typically lasting between 15 to 30 minutes. the average costs of the programme would fall between £28 and £71 per person in 2007/08 prices. guided selfhelp would be a cost-effective intervention for subthreshold or mild to moderate depression. It is difficult to assess whether. The aerobic forms of 50In the previous guideline the term ‘exercise’ was used. was used as an example for costing purposes. this intervention is likely to be delivered by a low-intensity therapy worker (essentially a paraprofessional) on the Agenda for Change (AfC) Band 5 salary scale. the average cost of the programme would range between £42 and £259 per person in 2007/08 prices. the booklet A Recovery Programme for Depression by Lovell and Richards (2008).Low-intensity psychosocial interventions The clinical evidence review above described interventions consisting of three to ten sessions (typically between four and six sessions) which were of limited duration over a 9.3. This cost includes salary.to 12-week period. 2009). it is assumed that the resources required to deliver the programme would be identical. based on these health service costs.00. If guided self-help were delivered on a group basis. 7. The total cost of individual or group-based guided self-help consists of the cost of staff plus the written self-help manual. Based on GDG opinion. Based on the estimated staff time associated with delivering an individual guided selfhelp programme as described above and the cost of an AfC Band 5 post (using the community mental health nurse costing).1 PHYSICAL ACTIVITY PROGRAMMES50 Introduction The effect of physical activity on mental health has been the subject of research for several decades.

Data were available to compare physical activity with a non-physical activity control.2 Studies considered for all comparisons51 In total.3. Twenty-five studies were included in the review. 2000). In the developed world. Physical activity may be divided into aerobic forms (training of cardio-respiratory capacity) and anaerobic forms (training of muscular strength/endurance and flexibility/co-ordination/relaxation) (American College of Sports Medicine. nine (Bosscher1993. McNeil1991.. 1980). Klein1985. trials not involving a physical activity intervention. Guidelines for physical activity referral schemes have been laid down by the Department of Health (2001..Low-intensity psychosocial interventions physical activity. 191 . and different kinds of physical activity. or trials not including participants with depression. Since there was a wide range of types of physical activity in the included studies. Principal reasons for exclusion included trials not being RCTs. resistance training). of which 25 were included and 32 were excluded. 1990. physical activity was defined as a structured physical activity with a recommended frequency. Singh1997. Herman2002. many of which include depression as a referral criterion. Combined data are reported here since an initial review of the evidence showed there was little difference between aerobic and non-aerobic physical activity. McCann1984. Definition For the purposes of the guideline. get positive feedback from other people and an increased sense of self-worth. running) and non-aerobic (for example. regular physical activity is seen as a virtue. have been most frequently investigated. 1994). In addition to the type of physical activity. as a result. In the past decade ‘exercise on prescription’ schemes have become popular in primary care in the UK (Biddle et al. It can be undertaken individually or in a group. papers not reporting outcome data. intensity and duration when used as a treatment for depression. Physical activity may act as a diversion from negative thoughts and the mastery of a new skill may be important (Lepore. Of these.. 59 RCTs were found. and physical activity may have physiological effects such as changes in endorphin and monoamine concentrations (Thoren et al. References for studies from the previous guideline are in Appendix 18. duration and intensity should be described. Several plausible mechanisms for how physical activity affects depression have been proposed. Mead et al. psychotherapy. Fremont1987. Greist1979. Social contact may be an important benefit. 7. 2008). Veale1992) were also included in the previous guideline. various combination treatments. especially jogging or running. waitlist or pill placebo. There were insufficient studies to 51Throughout the following sections on physical activity programmes. 1994). the frequency. Leith. 1997.. the depressed patient who takes regular physical activity may. study IDs in title case refer to studies included in the previous guideline and study IDs in capital letters refer to studies found and included in this guideline update. pharmacotherapy. the GDG divided these into aerobic (for example. Mynors-Wallis et al.

and so this factor was also included in the analysis. dichotomous efficacy outcomes were not extracted since these were reported by a relatively small number of studies. and some combination strategies (including physical activity plus light therapy) compared with no physical activity control or physical activity alone.Low-intensity psychosocial interventions look at specific types of activity separately. respectively. The following comparisons are not reported here: physical activity compared with other types of exercise.3 Physical activity versus no physical activity control. which also includes details of excluded studies. The GDG considered supervision to be an important factor in the success of physical activity programmes. whereas continuous outcomes were more widely reported. Summary study characteristics of the included studies are in Table 24 with full details in Appendix 17b. 192 . 7. The studies described below focus on the comparisons where substantial data were available. Clinical evidence Evidence from the important outcomes and overall quality of evidence are presented in Table 25 and Table 26. A sub-analysis was then carried out looking at group and individual physical activity separately.3. Clinical evidence summary Physical activity was more effective in reducing subthreshold symptoms and mild depressive symptoms than no physical activity control. The full evidence profiles and associated forest plots can be found in Appendix 16b and Appendix 19b. pill placebo and waitlist Studies considered Seventeen studies compared physical activity with no physical activity control. Since there was a large amount of data to report. although the effect was reduced at follow-up (see Table 25 and Table 26). The review team initially analysed the data combining group and individual physical activity and compared it with relevant control groups. Because of the large number of summary study characteristics and summary evidence profile tables for physical activity. a brief clinical evidence summary follows each of the summary evidence profile tables.

4) Diagnosis (average baseline score) 193 (1) BDI 14 (BDI 25. of participants) (1) BUTLER 2008* (2) MATHER 2002 (3) Singh1997 (4) SINGH1997A (5) SINGH2005D* (1) DUNN2005 (2) HOFFMAN 2008† (3) KNUBBEN 2007 (4) McNeil1991 (1) BLUMENTHAL2007† (2) HOFFMAN 2008† 5 (326) Study IDs (1) BERLIN 2003* (2) BLUMENTHAL2007† (3) HABOUSH 2006 (4) McCann1984 (5) TSANG2006 (6) Veale1992 (1) 46/74/50 (2) 86/69/65 (3) 32/63/71 (4) 32/53/71 (5) 60/55/69 (1) 80/75/36 (2) 202/76/52 (3) 38/55/50 (4) 30/?/73 (1) 202/76/52 (2) 202/76/52 (1) SIMS2006 N/% female/ mean age (1) 55/55/40 (2) 202/76/52 (3) 20/65/69 (4) 47/100/Not reported (5) 82/81/82 (6) 124/64/36 (1) 38/55/74 Low-intensity psychosocial interventions (1) GDS 11 (GDS 12. RCTs (No.Table 24: Summary of study characteristics of RCTs of physical activity versus no physical activity control.4) (2) MDD/ subthreshold (1) MDD/ subthreshold depressive symptoms/ dysthymia (BDI 30. placebo and waitlist Individual Supervised non-aerobic 5 (252) 4 (350) – 2 (404) 1 (38) Supervised aerobic Supervised non-aerobic Unsupervised aerobic Unsupervised non-aerobic Individual Group Supervised aerobic No.5) Continued .0) (2) MDD/ subthreshold depressive symptoms/ (1) Dysthymia/ depression/ subthreshold depressive symptoms/mild depressive (1) Moderate depressive episode (HRSD 19.

4) (4) BDI 12 and 24 (not reported) (2) MDD/ subthreshold depressive symptoms/ dysthymia (not reported) Supervised aerobic Supervised non-aerobic Unsupervised aerobic Individual Unsupervised non-aerobic 194 Group Low-intensity psychosocial interventions Supervised aerobic dysthymia (BDI 30.3) (5) Major depression/MDD/ subthreshold depressive symptoms/dysthymia (HRSD 19.1) depressive symptoms/ dysthymia (not reported) (3) Moderate depressive episode (CES-D 38.5) (3) HRSD 10 (HRSD 18.9) (4) Major depression/ subthreshold depressive symptoms/ dysthymia (BDI 19.8) (6) CIS 17 (not reported) .1) (3) MDD/ subthreshold depressive symptoms/ dysthymia (BDI 19.13) (4) BDI 11 (5) Features of depression/ depression/ dysthymia (GDS 5.Table 24: (Continued) Individual Supervised non-aerobic episode (not reported) (2) Mood disorder (HRSD 17.

or low-intensity resistance training (1) Treadmill/biking at different intensities/frequency (2) Aerobics (3) Treadmill (4) Walking (1) Walking/ jogging (2) Aerobics (1) Resistance training Control (1) No treatment (2) Placebo pill (3) Waitlist (4) No treatment (5) Newspaper reading (6) No treatment (1) Not reported (2) 10 weeks (3) 20 weeks (4) 10 weeks (5) 8 weeks (1) Not reported (2) 34 weeks (3) 6 weeks (4)–(5) Not reported (1)–(4) Not reported (1) 12 weeks (2) 16 weeks (3) 10 days (4) 6 weeks Length of treatment (1) 72 weeks (2) 16 weeks (3) 12 weeks (4) 10 weeks (5) 16 weeks (6) 12 weeks (1)–(2) 16 weeks (1) 10 weeks Low-intensity psychosocial interventions Follow-up (1)–(2) Not reported (1) 6 months 195 (1)–(2) Not reported (3) 3 months (4) Not reported (5) 8 weeks (6) Not reported *3-arm trial. .Physical activity (1) Water aerobics (2) Walking/ jogging (3) Ballroom dancing (4) Aerobic exercise (5) Qigong (6) Running (1)–(4) Health education (5) GP care (1) Stretching (2) Placebo pill (3) Stretching (4) Waitlist (1)–(2) Placebo pill (1) Advice control (1) Yoga (2)–(4) Resistance training (5) High. †4-arm trial.

24 (CI –0.41) 24 weeks: SMD 0.42 (CI –0.97) 34–36 weeks: SMD –0.67 to 0.01 PA 05.38 (CI –0.1 (CI –0. n Moderate Low Low K 7. n 82 34 weeks: K 1.6 to 0. n K 1.9% versus 12.72 to 3.29) 5.19 to –0.06 (CI –1.8) – – Not reported Not reported Follow-up Unsupervised Follow-up 196 Moderate K PA 01. n 86 PA 04. participants Forest plot number Low-intensity psychosocial interventions Self-rated mean depression change scores at endpoint Quality Number of studies. n 26 K 1. n 113 – – – SMD 0.15 (CI –0.7%) Low K PA 06.74 (CI –1.01) (17.01 4 weeks: SMD –1.01 RR 1. n 32 PA 03.Table 25: Summary evidence profile for physical activity versus no physical activity control Supervised SMD –1.08) 8 weeks: SMD –1. n 82 8 weeks: K 1.09 to –1.12 to –0.01 Not reported – 3.01 Not reported – – – Clinician-rated mean depression scores at endpoint Quality Number of studies.37 to 1. n 195 – – – SMD 0.01 Not reported – – – PA 05.59) 34 weeks: SMD –0.58 (CI –2.67 to 0. participants Forest plot number .21) 24 weeks: K 1.10 SMD –0. participants Forest plot number Leaving treatment early for any reason Quality Number of studies.53 to –0.75 to –0.47 (CI 0.26 (CI –2. n 23 34–36 weeks: K 2.01) 24 weeks: Low 34–36 weeks: High 213 PA 02.18) 4 weeks: Moderate 8 weeks: Moderate 34 weeks: Low 368 4 weeks: K 1.

02 Leaving treatment RR 0.3. n 22 Low K 1.02 7. K participants Forest plot number PA 07.4 Physical activity versus antidepressants Studies considered Three studies compared physical activity with sertraline.49 SMD –0.06 to 0.23) reason Quality Moderate 2. Clinical evidence Evidence from the important outcomes and overall quality of evidence are presented in Table 28. 197 .67 to 0.36) (CI –1. which also includes details of excluded studies.34 mean depression (CI –0. These studies have been classified based on whether physical activity was supervised or not and whether physical activity was conducted in groups or individually.Low-intensity psychosocial interventions Table 26: Summary evidence profile for physical activity versus control (pill placebo or waitlist) Supervised physical activity versus pill placebo Unsupervised physical activity versus pill placebo Supervised physical activity versus waitlist Supervised physical activity versus waitlist at follow-up (12 weeks) Clinician-rated SMD –0.24 to 0. n 100 Low K 1. Summary study characteristics of the included studies are in Table 27.2 (CI 0. K participants Forest plot number PA 09. with full details in Appendix 17b.64 early for any (CI 0. n 102 Low K 1.27) (CI –1. Clinical evidence summary The data comparing physical activity with sertraline were largely inconclusive being drawn from only three small studies.03 Not reported – – – – Number of studies. n 19 Number of studies.57) scores at endpoint Quality Low 1. n 102 PA 10.33 to 1.27 SMD –0.50 to 0.02 PA 09.65) Moderate K 1.12 SMD –0. n 170 PA 08. respectively.35 to 0. although there was some evidence suggesting that unsupervised physical activity may be more effective than antidepressants in reducing symptoms in subthreshold and mild depression.03 Not reported PA 11.02 RR 0. People taking antidepressants were more likely to leave treatment early because of side effects (see Table 28). The full evidence profiles and associated forest plots can be found in Appendix 16b and Appendix 19b.12) (CI –0.

5) Physical activity (1)–(2) Walking/jogging Control (1)–(2) Sertraline Length of treatment (1)–(2) 16 weeks Follow-up (1) Not reported (2) 24 weeks *4-armed trial. RCTs (No. . of participants) 2 (358) Study IDs (1) BLUMENTHAL 2007* (2) Herman2002 Low-intensity psychosocial interventions N/% female/ mean age (1) 202/76/52 (2) 156/73/57 Diagnosis (average baseline score) (1) MDD/subthreshold depressive symptoms/ dysthymia (BDI 30.198 Individual Unsupervised aerobic 1 (202) (1) BLUMENTHAL 2007* (1) 202/76/52 (1) MDD/subthreshold depressive symptoms/ dysthymia (BDI 30.5) (2) MDD/subthreshold depressive symptoms/ dysthymia (BDI 22.5) (1) 202/76/52 (1) HOFFMAN2008* 1 (202) 1 (202) (1) HOFFMAN2008* Supervised aerobic Unsupervised aerobic (1) 202/76/52 (1) MDD/subthreshold (1) MDD/subthreshold depressive symptoms/ depressive symptoms/ dysthymia (not reported) dysthymia (not reported) (1) Walking/jogging (1) Sertraline (1) 16 weeks (1) Not reported (1) Aerobics (1) Sertraline (1) 16 weeks (1) Not reported (1) Aerobics (1) Sertraline (1) 16 weeks (1) Not reported Table 27: Summary study characteristics for physical activity versus antidepressants Group Supervised aerobic No.

participants Forest plot number SMD –0.19 scores at endpoint (CI –0. Summary study characteristics of the included studies are in Table 29 with full details in Appendix 17b. participants Forest plot number Low K 1.03 (CI –1. n 149 PA 15.61) Moderate K 1.20) Quality Number of studies. n 102 PA 16. n 102 PA 12.2%) Moderate K 2.41 (CI 1.79 to 0. participants Forest plot number Leaving treatment early due to side effects Quality Number of studies. n 101 – – RR 0.04 PA 14.11 to 1.2% versus 6.04 RR 7.78) Low K 1. n 201 K 1.04 RR 1.4 (CI 0. n 201 Unsupervised aerobic SMD –1.77 (CI 0.23) (19.04 PA 16.44 to –0. 199 .45) (5.1% versus 14.Low-intensity psychosocial interventions Table 28: Summary evidence profile for physical activity compared with antidepressants Supervised aerobic Clinician-rated mean depression scores at endpoint Quality Number of studies.7% versus 14.04 Not reported – Self-rated mean depression SMD –0.58 to 0.4%) Moderate K 2.4 to 39.9) (23.04 RR 2. n 102 PA 15. participants Forest plot number Leaving treatment early for any reason Quality Number of studies.5 Physical activity versus psychosocial and psychological interventions Studies considered Four studies compared physical activity with a psychosocial or psychological intervention.87 to 2.3.3 to 25.3%) Low PA 13.75 (CI –1. which also includes details of excluded studies.28) Low K 2.59 (CI 0.04 7.

Table 29: Summary study characteristics for physical activity versus psychosocial and psychological interventions Individual Supervised non-aerobic 1 (46) (1) BUTLER2008* (1) Klein1985 (2) McNeil1991 (1) 74/72/30 (2) 30/?/73 (1) Major/subthreshold depressive symptoms (not reported) (2) BDI 12 and 24 (not reported) (1) Running (2) Walking (1) Group therapy (2) Social contact (1) 12 weeks (2) 6 weeks (1) Not reported (1) 1. RCTs (No. 3 and 9 months (2) Not reported (1) 46/74/50 (1) Dysthymia/depression/ subthreshold depressive symptoms/dysthymia/mild depressive episode (not reported) (1) Yoga (1) Hypnosis 2 (104) – – – – Supervised aerobic Supervised non-aerobic 200 – – (1) Not reported – – Group Supervised aerobic No. of participants) 2 (89) Study IDs (1) Fremont1987 (2) Greist1979 Low-intensity psychosocial interventions N/% female/mean age (1) 61/74/unclear (2) 28/54/24 Diagnosis (average baseline score) (1) BDI 9–30 (not reported) (2) RDC for minor depression and SCL-90 depression cluster score at 50th percentile or above Physical activity (1) Running (2) Running Control (1) Cognitive techniques (2) Time-limited psychotherapy Length of treatment (1) 10 weeks (2) 10 weeks Follow-up (1) 2 and 4 months (2) Not reported .

68 to 0.59 to 0.41 (CI –1. n 18 PA 19. n – – Forest plot number Leaving treatment early for any reason Low-intensity psychosocial interventions Quality 201 Number of studies.01 Not reported 26 36 weeks: SMD –0.94 to 0.80 (CI –0.18 to 0.23 (CI –0.79 to 0. participants Forest plot number Self-rated mean depression scores at endpoint Quality Moderate – Number of studies.58) (20% versus 16.04 to 1.2 (CI 0.17 (CI –0.05 Not reported – 1. n 26 34 weeks: K 1. participants PA 18.Table 30: Summary evidence profile for physical activity versus psychosocial and psychological interventions Supervised aerobic Not reported Not reported SMD 0.7%) Low K PA 20.64) Low K PA17 05. n 31 16 weeks: K 1.62) 16 weeks: SMD –0.01 Not reported 1.63 (CI –1. n 16 – – Not reported – – – Not reported – – – – – – – Quality Number of studies.14 to 10.33) 8 weeks: Low 16 weeks: Low 34 weeks: Low 79 8 weeks: K 1.05 K 3.21) 8 weeks: SMD –0.09 (CI –0. n 24 K Low 1.05 RR 1. n PA17 05.60) Follow-up Supervised non-aerobic Follow-up Clinician-rated mean depression scores at endpoint – – – SMD –0. participants Forest plot number .37) 34 weeks: SMD –0.

5) (average baseline score) (2) MDD/subthreshold depressive symptoms/dysthymia (HAM-D 19. which also includes details of excluded studies. RCTs (No.6 Physical activity plus antidepressants versus antidepressants Studies considered Two studies compared physical activity and antidepressants versus antidepressants. Table 31: Summary study characteristics for physical activity antidepressants versus antidepressants Group Supervised aerobic physical activity antidepressants versus antidepressants No.3. respectively.9) Physical activity Control Length of treatment Follow-up (1) Running (2) Running sertraline antidepressant (range of drugs used) (1) Sertraline (2) Combination antidepressants (range of drugs used) (1) 16 weeks (2) 32 weeks (1) 24 weeks (2) Not reported 202 . Clinical evidence summary The data from four studies comparing physical activity with a range of psychosocial and psychological interventions are insufficient to determine a clear picture of the relative effectiveness of physical activity (see Table 30). 7.Low-intensity psychosocial interventions Clinical evidence Evidence from the important outcomes and overall quality of evidence are presented in Table 30. Summary study characteristics of the included studies are in Table 31 with full details in Appendix 17b. The full evidence profiles and associated forest plots can be found in Appendix 16b and Appendix 19b. of participants) Study IDs N/% female/mean age 2 (186) (1) Herman2002 (2) PILU2007 (1) 156/73/57 (2) 30/100/unclear Diagnosis (1) Major depressive disorder (BDI 22.

n 30 antidepressants Supervised aerobic physical activity antidepressant versus antidepressant SMD –0.37 (CI 0.23) Moderate K 1. participants Forest plot number SMD –1.47) Moderate K 1.26) (20% versus 14.85 to –0. The full evidence profiles and associated forest plots can be found in Appendix 16b and Appendix 19b.08 (CI –0.47 to 0. respectively. participants Forest plot number Leaving treatment early because of side effects Quality Number of studies.87 (CI 0.09 PA 21. n 103 PA 21. n 103 – – – PA 23. participants Forest plot number – – – Not reported PA22 09.6%) Low K 1.09 203 .Low-intensity psychosocial interventions Clinical evidence Evidence from the important outcomes and overall quality of evidence are presented in Table 32.58 to 3. participants Forest plot number Leaving treatment early for any reason Quality Number of studies.1% versus 10.09 RR 0.31) Moderate K 1.04 (CI –1.83) Not reported (9. n 103 Self-rated mean Not reported depression scores at endpoint Quality Number of studies. Table 32: Summary evidence profile for physical activity versus antidepressants Supervised aerobic physical activity antidepressant versus combination antidepressants Clinician-rated mean depression scores at endpoint Quality Number of studies.4%) Low K 1. n 103 – – – PA 24.27 to 2.31 to 0.09 SMD 0.02 RR 1.08 (CI –0.

There appeared to be no difference between combination treatment versus a single antidepressant. by looking at comparisons between group-based physical activity programmes compared with no physical activity control. respectively. Clinical evidence summary The intensity of many of the physical activity programmes (for example. a subgroup analysis of group programmes was undertaken.7 Sub-analysis of group-based versus individual physical activity programmes Studies considered A sub-analysis was conducted to examine the indirect effectiveness of group-based physical activity programmes in comparison with individual physical activity. 204 . The full evidence profiles and associated forest plots can be found in Appendix 16b and Appendix 19b. Clinical evidence Evidence from the important outcomes and overall quality of evidence are presented in Table 34 and Table 35. it is difficult to draw any conclusions (see Table 32). based on the previous results of physical activity. there was no clear benefit of individual over group-based physical activity. As there was only one study in each comparison. three supervised sessions per week over a 12-week period) raises questions about the cost effectiveness of individual physical activity programmes. which also includes details of excluded studies. Furthermore. 7. To address this.Low-intensity psychosocial interventions Clinical evidence summary Physical activity plus an antidepressant more effectively reduced depression scores than a combination of two antidepressants in one study. Summary study characteristics of the included studies are in Table 33 with full details in Appendix 17b. The GDG decided to carry out this indirect comparison given that the cost of individual physical activity is considerably greater than group-based. and also looking at comparisons between individual physical activity and no physical activity control.3. This was performed indirectly. An indirect comparison can be made by looking at Table 34 and Table 35.

of RCTs (No.4) (2) Moderate depressive episode (CES-D 38.4) 205 (1) BDI 14 (BDI 25.4) Continued . of participants) 5 (326) Study IDs (1) BERLIN2003* (2) BLUMENTHAL2007† (3) HABOUSH2006 (4) McCann1984 (5) TSANG2006 (6) Veale1992 N/% female/Mean age (1) 55/55/40 (2) 202/76/52 (3) 20/65/69 (4) 47/100/Not reported (5) 82/81/82 (6) 124/64/36 (1) 38/55/74 Low-intensity psychosocial interventions Diagnosis (average baseline scores) (1) GDS 11 (GDS 12.5) (3) HRSD 10 (HRSD 18.1) (1) Moderate depressive episode (HRSD 19. placebo and waitlist Supervised non-aerobic 5 (252) (1) BUTLER2008* (2) MATHER2002 (3) Singh1997 (4) SINGH1997A (5) SINGH2005D* (1) 46/74/50 (2) 86/69/65 (3) 32/63/71 (4) 32/53/71 (5) 60/55/69 (1) 80/75/36 (2) 38/55/50 Supervised aerobic No.0) (2) MDD/subthreshold depressive symptoms/ dysthymia (BDI 30.Table 33: Summary study characteristics for group-based and individual physical activity programmes Individual physical activity versus no physical activity control Supervised aerobic 2 (118) (1) DUNN2005 (2) KNUBBEN2007 1 (38) (1) SIMS2006 Unsupervised non-aerobic Group physical activity versus no physical activity control.13) (1) Dysthymia/depression/ subthreshold depressive symptoms/mild depressive episode (not reported) (2) Mood disorder (HRSD 17.

52% no formal diagnosis.1) (1) Yoga (2)–(4) Resistance training (5) High. placebo and waitlist Low-intensity psychosocial interventions Supervised aerobic (4) BDI 11 (5) 45% depression. 3% dysthymia (6) CIS 17 (not reported) Physical activity (1) Water aerobics (2) Walking/jogging (3) Ballroom dancing (4) Aerobic exercise (5) Qigong (6) Running .or low-intensity resistance training (1) Treadmill/biking at different intensities/frequency (2) Treadmill (1) Resistance training Group physical activity versus no physical activity control.9) (4) Major depression/ subthreshold depressive symptoms/dysthymia (BDI 19.206 Table 33: (Continued) Individual physical activity versus no physical activity control Supervised aerobic Unsupervised non-aerobic Supervised non-aerobic (3) MDD/subthreshold depressive symptoms/ dysthymia (BDI 19.3) (5) Major depression/MDD/ subthreshold depressive symptoms/dysthymia (HRSD 19.

207 . †4-arm trial.Comparator (1) No treatment (2) Placebo pill (3) Waitlist (4) No treatment (5) Newspaper reading (6) No treatment (1) Not reported (2)–(4) 10 weeks (5) 8 weeks (1) 12 weeks (2) 10 days (1) 10 weeks (1)–(4) Health education (5) GP care (1)–(2) Stretching (1) Advice control Length of treatment (1) 72 weeks (2) 16 weeks (3) 12 weeks (4) 10 weeks (5) 16 weeks (6) 12 weeks (1) Not reported (2) 34 weeks (3) 6 weeks (4)–(5) Not reported (1)–(2) Not reported Follow-up (1)–(2) Not reported (3) 3 months (4) Not reported (5) 8 weeks (6) Not reported (1) 6 months Low-intensity psychosocial interventions *3-arm trial.

n 23.15 (CI –0.09 to –1.02 SMD –0.94 (CI –1.53 to –0.02.02.67 to 0. 19.06 (CI –1. n 86 Table 34: Group-based physical activity programmes versus no physical activity control Clinician-rated mean depression scores at endpoint Low-intensity psychosocial interventions Quality Number of studies. 113 PA26. n 82. n 147 K K 1.24) 34 weeks: SMD –0. n 2.29 to –0.24 (CI –0.59) Moderate Moderate 2. 19.208 Supervised aerobic Not reported Not reported SMD –0. 19.01.01 High Moderate K 4. 82 K 4.38 (CI –0.58 (CI –2.01) Low High 183 K K 1. n – Moderate – SMD –0.77 (CI –1.97) 34–36 weeks: SMD–0. n 1.01 PA26. participants K . n 183 Moderate K 1. participants – Forest plot number Self-rated mean depression scores at endpoint Quality Number of studies.75 to –0.67 to 0.08 to –0.18) – – – 4 weeks: SMD –1.45) Follow-up Supervised non-aerobic Follow-up 24 weeks: SMD 0.84 to –0.08) 8 weeks: SMD –1.59) PA25.54 (CI –0.

26 to 98.20 to 20.0 (CI 0.7%) Moderate 2.19 Self-rated depression change scores at endpoint Quality Number of studies. participants – – Low-intensity psychosocial interventions Forest plot number 209 . n 40 – – Not reported 39 – – – – – – Not reported Not reported Not reported PA 27.26 to 0.0 (CI 0.56 to 2. participants K PA 28.19 PA 27.Forest plot number SMD –0.33) (10% versus 5%) Low K 1. participants K PA 30.79) (20.19 PA 29.19 RR 5.00) (10% versus 0%) Low K 1.3% versus 15.19 RR 1.19 40 Leaving treatment early because of side effects – – – – Not reported Quality – – – Number of studies.24 (CI 0.19 PA 29. n – Not reported 115 – – Forest plot number Leaving treatment early for any reason Quality – – – Number of studies. n PA 31.61 (CI –1. n – Not reported RR 2.19 Not reported Forest plot number PA 30.03) Low 1.

–0. n 26 Low K 1. 95% CI –2. n 30 Unsupervised non-aerobic Not reported Follow-up Not reported – – – – – – PA 32.20 PA 33.45) and at 34 to 36 weeks’ follow-up (SMD –0. 95% CI –1. and is a challenging dataset to interpret. the evidence indicates that supervised non-aerobic group physical activity has a significant effect in patients with largely subthreshold depression in reducing clinician-reported depression scores at endpoint (SMD –0.75.54.08.16 (CI –1.24. 95% CI –1.24).59).58.60 to 0. participants Forest plot number Self-rated mean depression scores at endpoint Quality Number of studies.000 participants. –0.67.20 SMD –0.54 to –0. 95% CI –0.42 24 weeks: SMD (CI –1.80) Moderate K 1. –0. which comprised mixed groups of patients with mild major depression.94.18).3.21) 0. the evidence indicates that group-based physical activity is effective in the treatment of subthreshold and mild depression.59).20 PA 34.53. dysthymia and subthreshold depressive symptoms. See Table 34 and Table 35.08) and at 8 weeks (SMD –1.37) Moderate K 1. at 4 weeks’ follow-up (SMD –1.8 Clinical evidence summary for physical activity – all comparisons The evidence is presented for a relatively large dataset of 25 trials and over 2. Supervised non-aerobic group physical activity had a positive effect in reducing self-rated depression scores at endpoint (SMD –0. 95% CI –1.20) (CI –0. and at 34 weeks’ follow-up (SMD –0. participants Forest plot number SMD –1.Low-intensity psychosocial interventions Table 35: Individual physical activity versus no physical activity control Supervised aerobic Clinician-rated mean depression scores at endpoint Quality Number of studies. –0. n 32 PA 33.29.84. Supervised aerobic group physical activity had a beneficial effect in reducing self-rated depression scores at endpoint (SMD –0.01).38.10 (CI –0. 0.06. The participants who were included in the trials in this review who met criteria 210 . When compared with no physical activity controls. 7.09. n 38 Low K 1.87 SMD 0.94 to –0.37 to 1.77. –0. This stems from a number of factors including the variation in the populations. –1. 95% CI –0. 95% CI –0. 20 Overall.

the nature of the physical activity interventions was also very varied. In addition. In the absence of any clear and direct indication from the data of benefits for a particular kind of physical activity (for example. The data comparing physical activity with antidepressants suggests no significant differences. however.58.06.28 and for self-rated scores: SMD –0. and.67.87.9 Health economic evidence and considerations No evidence on the cost effectiveness of structured physical activity programmes for people with subthreshold depressive symptoms or mild to moderate depression was 211 . 95% CI –1. and at 34–36 weeks: SMD –0.38. there is insufficient evidence to identify any differential effect. the data suggest that physical activity is more effective in reducing depressive symptoms than a no physical activity control (clinician-rated scores: SMD –1. 95% CI –0. anaerobic versus aerobic).53.68.75. a benefit for group-based physical activity. The data for physical activity compared with psychosocial and psychological interventions for depression did not suggest any important differences.31. 95% CI –1. the CIs were wide (for clinician-rated scores: SMD –0. The CIs of this dataset were wide and there is insufficient evidence on which to make any clear conclusion.34).12.41. As expected. 95% CI –0. 95% CI –2. 0. people taking antidepressants were more likely to leave treatment early (RR 1. Taken together.83. 0. –0. 95% CI –0.9). The effectiveness for physical activity when compared with pill placebo came from only two studies. the GDG took the view that patient preference should be a significant factor in determining the nature of the physical activity.21). and for self-rated scores at 4 weeks: SMD –1. self-reported scores: –0.75. Some comparators were also potentially problematic with one study having combined antidepressants as the comparator in a population that included those with a diagnosis of dysthymia and subthreshold depressive symptoms (PILU2007). Despite these issues.15.19. the GDG considered the potentially limited effectiveness of individual physical activity (with a high level of contact: up to three sessions per week over a 10. as indeed were the comparators.59. 0. more specifically.23.08.59. –1. 7.01. 95% CI –2. therefore. 0. 2.79.Low-intensity psychosocial interventions for major depressive disorder were drawn predominantly from groups in the mild to moderate range of depression (mean baseline BDI scores between 18 and 25). –0.24. –0. although the effect was reduced at follow-up (clinician-rated scores at 24 weeks: SMD 0.3. 0. 95% CI 0. these studies suggest a benefit for physical activity in the treatment of subthreshold depressive symptoms and mild to moderate depression.26.09.to 12-week period). 95% CI –0. Physical activity also has the advantage of bringing other health gains beyond just improvement in depressive symptoms.97. There were further studies that compared the effectiveness of one type of physical activity with another type of physical activity or a combination of activities. –0. but again the results were difficult to interpret given the width of the CIs (for self-rated scores: SMD –0.18).67. 95% CI –1. In addition to the effectiveness of group-based physical activity. The results did not indicate a clear picture favouring any specific treatment (see Appendix 19b). 95% CI –0.58. and at 34 weeks: SMD –0. at 8 weeks: SMD –1.20).

This cost includes salary. It is likely that the sessions would be supervised by a physical activity facilitator (an NHS professional or paraprofessional with expertise in the area) who would be a recent graduate from an undergraduate or Masters’ level course.4 FROM EVIDENCE TO RECOMMENDATIONS – LOW-INTENSITY PSYCHOSOCIAL INTERVENTIONS A range of low-intensity interventions (CCBT. The relative cost-effectiveness of an individual or group-based physical activity programme also depends on the impact on downstream resource use and not just the service costs of delivering the interventions.142 per person in 2007/08 prices. 7.1. Therefore. It is difficult to assess whether. which is likely to be comparable to the salary scales of a community mental health nurse. The clinical evidence in the literature review described interventions delivered either individually or in structured groups under the supervision of a competent practitioner or physical activity facilitator. The unit cost of an AfC Band 5 community mental health nurse is £51 per hour of patient contact in 2007/08 prices (Curtis. However. guided self-help and group-based physical activity programmes) have been identified as being effective for subthreshold 212 . If a physical activity programme were delivered on a structured group basis. it is possible that this will be more cost effective than an individual programme for patients with subthreshold or mild to moderate depression. Details on the methods used for the systematic search of the economic literature are described in Chapter 3.to 14-week period. it is difficult to ascertain whether an individual physical activity programme is more or less clinically effective than a group-based programme. Based on the estimated staff time associated with delivering and supervising a physical activity programme as described above and the cost of a community mental health nurse. The programme would typically involve two to three sessions per week of 45 minutes’ to 1 hour’s duration over a 10. Section 3. a physical activity programme would be a cost-effective intervention for subthreshold or mild to moderate depression. The unit cost of a physical activity facilitator is not currently available. the average cost of a physical activity programme when delivered at an individual level would range between £765 to £2. salary on-costs. the average costs of the programme would fall between £128 to £428 per person in 2007/08 prices. Therefore. it is difficult to assess how these clinical improvements could be translated into overall improvements in patient HRQoL that could be used in a cost-effectiveness analysis. given the lower costs of delivering a structured group-based physical activity programme. it is assumed that such workers would be on AfC salary scales 4 or 5. it is assumed that the resources required to deliver the programme would be identical. overheads and capital overheads plus any qualification costs. However.6. The clinical evidence suggests that both individual and structured group physical activity interventions are effective in reducing symptoms of depression when compared with a no physical activity control. based on these health service costs. Based on the assumption of five to six people per group. 2009).Low-intensity psychosocial interventions identified by the systematic search of the health economics literature.

. No evidence on the cost effectiveness of either guided self-help or physical activity for subthreshold depressive symptoms or mild to moderate depression was identified in the systematic review of the health economic literature.1 For people with persistent subthreshold depressive symptoms or mild to moderate depression. Simple cost analyses combined with the limited clinical evidence suggests that guided self-help interventions may be cost effective compared with control treatments or treatment as usual and that the preferred mode of delivery for physical activity is in groups. be guided by the preference of people with depression and this is reflected in the recommendations. apart from the fact that both guided self-help and CCBT should be based on cognitive behavioural principles and that physical activity should be delivered in a group format. The data also did not support the view that any particular mode of delivery (for example.5.4.5. aerobic versus anaerobic physical activity. All interventions seem to require some form of support or supervision to be fully effective.. The GDG was also concerned that the effective delivery of the interventions may be compromised by differences in the style and content of delivery of the intervention and so has drawn on existing trial data to offer specific recommendations on the content of the interventions.1 in Clinical Guideline 91 [NICE.2 Individual guided self-help programmes based on the principles of CBT (and including behavioural activation and problem-solving techniques) 52This recommendation (and recommendation 1.1. internet versus desktop-based CCBT) for any low-intensity intervention had any specific advantage over another. 2006a]). Delivery of low-intensity psychosocial interventions 7. Kaltenthaler et al. There are few trials that allow for direct clinical or cost-effectiveness comparisons of any of the interventions.Low-intensity psychosocial interventions depressive symptoms and mild to moderate depression. Based on the health economic evidence a variety of CCBT packages were judged to be cost effective when compared with standard care or treatment as usual in the treatment of subthreshold depressive symptoms or mild to moderate depression (McCrone et al. 213 .1.5 RECOMMENDATIONS Low-intensity psychosocial interventions 7. As a result the GDG took the view that the decision as to which intervention to offer should. 2003. consider offering one or more of the following interventions.2. guided by the person’s preference: ● individual guided self-help based on the principles of cognitive behavioural therapy (CBT) ● computerised cognitive behavioural therapy (CCBT)52 ● a structured group physical activity programme. 2008). 2009c]) updates the recommendations on depression only in ‘Computerised cognitive behavioural therapy for depression and anxiety (review)’ (NICE technology appraisal 97 [NICE. 7. in significant part.

who typically facilitates the self-help programme and reviews progress and outcome ● consist of up to six to eight sessions (face-to-face and via telephone) normally taking place over 9 to 12 weeks. 214 . including: ● establishing regular sleep and wake times ● avoiding excess eating.5. Physical activity programmes for people with persistent subthreshold depressive symptoms or mild to moderate depression should: ● be delivered in groups with support from a competent practitioner ● consist typically of three sessions per week of moderate duration (45 minutes to 1 hour) over 10 to 14 weeks (average 12 weeks).Low-intensity psychosocial interventions for people with persistent subthreshold depressive symptoms or mild to moderate depression should: ● include the provision of written materials of an appropriate reading age (or alternative media to support access) ● be supported by a trained practitioner.5 Offer people with depression advice on sleep hygiene if needed. thought patterns and outcomes ● be supported by a trained practitioner.1. 53The evidence for this recommendation has not been updated since the previous guideline. including follow-up. who typically provides limited facilitation of the programme and reviews progress and outcome ● typically take place over 9 to 12 weeks.3 7.1. CCBT for people with persistent subthreshold depressive symptoms or mild to moderate depression should: ● be provided via a stand-alone computer-based or web-based programme ● include an explanation of the CBT model. and use thought-challenging and active monitoring of behaviour.1. smoking or drinking alcohol before sleep ● creating a proper environment for sleep ● taking regular physical exercise53.4 Sleep hygiene 7. 7.5. Any wording changes have been made for clarification only. encourage tasks between sessions. including follow-up.5.

. 2007). As with any psychological treatment.1. Scott et al. problem solving therapy and couples therapy54 share a common theoretical base. 2003.. people focus on negative views of themselves. Beck during the 1950s and was formalised into a treatment in the late 1970s (Beck et al. transference or offer interpretation as in psychodynamic psychotherapy. 1979). 2004a). the world and the future. through collaboration.. 215 . The cognitive model describes how.1 COGNITIVE BEHAVIOURAL THERAPIES Introduction Cognitive behavioural therapy (CBT) for depression was developed by Aaron T. 8.1 8. which Beck posited was the result of the operation of underlying cognitive schemas or beliefs. the person with depression learns to recognise his or her negative thinking patterns and to re-evaluate his or her thinking. Its original focus was on the styles of conscious thinking and reasoning of depressed people. 2000. The guideline refers to ‘cognitive behavioural therapies’ to indicate the evolution of CBT for depression over several decades. they are reviewed separately. 1999. behavioural activation. 54Five out of six of the included studies of couples therapy were based on a behavioural model. cognitive behavioural therapy is not static and has been evolving and changing.High-intensity psychological interventions 8 HIGH-INTENSITY PSYCHOLOGICAL INTERVENTIONS This section covers the high-intensity interventions that were identified in the searches for the guideline update and groups them according to the definitions developed for the previous guideline (NICE. 1997) to address underlying beliefs more directly. The therapy takes an educative approach where. This approach also requires people to practise re-evaluating their thoughts and new behaviours (called homework). The approach does not focus on unconscious conflicts. Watkins et al. There have been important elaborations on the techniques of therapy (Beck.. when depressed. Although to some degree cognitive behavioural therapies. which have been applied to particular presentations such as persistent residual depressive symptoms that leave people vulnerable to relapse (Paykel et al. Moore & Garland.

thoughts and feelings associated with depressive relapse. MBCT is intended to enable people to learn to become more aware of the bodily sensations. which has demonstrated efficacy in preventing depressive relapse/recurrence (Hollon et al. One. the manual used was Beck’s Cognitive Therapy of Depression (1979) which advocates 16 to 20 sessions for treatment and relapse prevention work. With 8 to 15 patients per group. and to relate constructively to these experiences. cognitive behavioural therapies were defined as discrete. structured psychological interventions. It consists of 12 sessions of 2 hours’ duration over 8 weeks with groups held twice weekly for the first 4 weeks.and 6month follow-up sessions are also held and booster sessions can be used to help prevent relapse. These strategies include improving social skills. 1989). MBCT is an 8-week group programme with each session lasting 2 hours. The groups are highly structured (Lewinsohn et al.. 1984.. beliefs and interpretations on current symptoms. 1990). feelings states and/or problem areas ● develops skills to identify. 2005. MBCT has the potential to help a large number of people. This approach has a strong psychoeducational component focused on teaching people techniques and strategies to cope with the problems that are assumed to be related to their depression. Lewinsohn et al. and relaxation training.. group-based skills training programme designed to enable patients to learn skills that prevent the recurrence of depression (Segal et al. increasing pleasant activities. Group cognitive behavioural therapy Trials looking at group CBT. 2002). with two group leaders.. 2003. beliefs and/or problem areas. 1986) and typically consist of six to ten adults. MBCT is a manualised. Mindfulness-based cognitive therapy Mindfulness-based cognitive therapy (MBCT) was developed with a specific focus on preventing relapse/recurrence of depression (Segal et al. beliefs and interpretations related to the target symptoms/problems ● learns a repertoire of coping skills appropriate to the target thoughts.High-intensity psychological interventions Definition For the purpose of this review. were also included. It is based on theoretical and empirical work demonstrating that depressive relapse is associated with the reinstatement of automatic modes of thinking. a programme with proven efficacy in ameliorating distress in people with chronic disease (Baer. 2005). and four follow-up sessions in the year after the end of therapy. and CBT for acute depression (Beck et al. In most individual trials of CBT. monitor and then counteract problematic thoughts... 2002). Kabat-Zinn. derived from the cognitive behavioural model of affective disorders and where the patient: ● works collaboratively with the therapist to identify the types and effects of thoughts. addressing negative thinking. It is derived from mindfulness-based stress reduction. 1979). Lewinsohn et al. time-limited.. feeling and behaving that are counter-productive in contributing to and 216 . which predominantly uses the ‘Coping With Depression’ approach (Kuehner.

Table 40 and Table 41...High-intensity psychological interventions maintaining depressive relapse and recurrence (for example. 24 studies were found in the update search and 22 were also reported in the previous guideline. patients develop an ‘action plan’ that sets out strategies for responding when they become aware of early warning signs of relapse/recurrence (Williams. each study considered for review is referred to by a ‘study ID’ made up of first author and publication date (unless a study is in press or only submitted for publication. when first author only is used).1. self-critical thinking and avoidance) (Lau et al.2 Studies considered55 In total.. Study IDs in title case refer to studies included in the previous guideline and study IDs in capital letters refer to studies found and included in this guideline update. respectively. 68 studies were identified. Table 37 and Table 38. 22 trials were excluded in this update search. of which 46 RCTs were included. with full details in Appendix 17b. by learning that ‘thoughts are not facts’). References for studies from the previous guideline are in Appendix 18 and references for studies for the update are in Appendix 17b. 8. step out of them and respond in healthier ways by intentionally moving into a mode in which they ‘de-centre’ from negative thoughts and feelings (for example. J.3 Clinical evidence for cognitive behavioural therapies Evidence from the important outcomes and overall quality of evidence are presented in Table 39.M. Furthermore. or more than 50% of participants dropped out of the study. Participants learn to recognise these ‘automatic pilot’ modes. 2008). et al. accept difficulties using a stance of selfcompassion and use bodily awareness to ground and transform experience. 8. The main reasons for exclusion were: trials included populations that were not diagnosed with depression. 217 . The full evidence profiles and associated forest plots can be found in Appendix 16b and Appendix 19b. 55Here and elsewhere in the guideline. In the latter stages of the course. which also includes details of excluded studies.1. authors replaced dropouts. 2004). Summary study characteristics of the included studies are presented in Table 36.

218 CBT versus comparator CBT versus rational emotive behavioural therapy 1 RCT (180) (1) DAVID2008 (1) DIMIDJIAN 2006 (2) Gallagher1982 (3) JACOBSON 1996 3 RCTs (216) 4 RCTs (502) 2 RCTs (383) CBT versus behavioural activation CBT versus interpersonal therapy (IPT) CBT versus short-term psychodynamic psychotherapy CBT (primary care) versus GP care 3 RCTs (202) (1) Freeman2002 (2) Scott1992 (3) Scott1997 6 RCTs (549) (1) Beach1992 (2) DERUBEIS 2005 (3) DIMIDJIAN 2006 (4) Elkin1989 (5) Jarrett1999 (6) Selmi1990 (1) Elkin1989 (1) Gallagher(2) Freeman2002 Th1994 (3) LUTY2007 (2) Shapiro1994 (4) MARSHALL 2008 (1) 45/100 (2) 141/59 (3) 159/66 (1) 113/66 (1) 159/66 (2) 23/77 (3) 110/72 (1) 168/70 (2) 96/61 (3)–(4) 70/69 (1) 60/92 (2) 61/52 (1) 96/61 (2) 91/75 (3) 32/67 Table 36: Summary study characteristics of cognitive behavioural therapies CBT versus antidepressants No. trials (total participants) 18 RCTs (1982) High-intensity psychological interventions Study IDs (1) BAGBY2008 (2) Blackburn1981 (3) Blackburn1997 (4) DAVID2008 (5) DERUBEIS2005 (6) DIMIDJIAN2006 (7) Elkin1989 (8) Hautzinger (in-pats) (9) Hautzinger1996 (10) Jarrett1999 (11) Keller2000 (12) LAIDLAW2008 (13) MARSHALL2008 (14) Miranda2003 (15) Murphy1984 (16) Murphy1995 (17) Scott1997 (18) Thompson2001 N/% female (1) 175/63 (2) Unextractable (3) 31/58 .

(4) 113/66 (5) 141/59 (6) 159/66 (7) 168/70 (8) 113/63 (9) 113/63 (10) 51/71 (11) 445/65 (12) 29/73 (13) 70/69 (14) 267/100 (15) 52/74 (16) 37/70 (17) 32/67 (18) 67/67 (1) 39 (2)–(3) 40 (4) 35 (5) 40 (6) 28 (1) 37 (1) 40 (2) 68 (3) 30 (1) 35 (2) 37 (3) 35 (4) No information (1) 62 (2) 41 (1) 37 (2) 32 (3) 41 (4) 168/70 (5) 48/67 (6) 23/64 Mean age (1) 42 (2) 43 (3) 40 (4) 37 (5)–(6) 40 (7) 35 (8)–(9) 40 (10) 39 (11) 43 (12) 76 (12) No information (13) 33 (14) 29 (15) 41 (16) 39 (17) 67 (18) 67 High-intensity psychological interventions 219 Continued .

20% anxiety disorder and 34% comorbid diagnosis (3)–(4) 100% MDD (1) 100% RDC depression (2) Major depressive episode CBT versus behavioural activation CBT versus IPT CBT versus short-term psychodynamic psychotherapy CBT (primary care) versus GP care (1) 46% mood disorder. 20% anxiety disorder and 34 comorbid diagnoses (2)–(3) 100% MDD 220 (1) 91% depressive (1) 100% MDD episode. 9% dysthymia (2)–(5) 100% MDD (6) 100% major or intermittent depression. 20% dysthymia (10)–(18) 100% MDD Comparator (1) SSRIs (2) Amitriptyline/ clomipramine (3) Antidepressants of GP’s choice (4) Fluoxetine (40–60 mg/day) (5) Paroxetine (38 mg/day) (6) Paroxetine (7) Imipramine (200–300 mg/day) (8)–(9) Amitriptyline (150 mg/day) (10) Phenelzine (0. usual treatment CBT versus antidepressants High-intensity psychological interventions Diagnosis (1) 100% major depressive episode (2) 100% MDD (3) 100% episode of depression (4)–(7) 100% MDD (8)–(9) 80% MDD. or subthreshold depressive symptoms (1) Waitlist (2)–(3) Placebo (4) Placebo clinical management (5) Placebo (6) Waitlist (1) Antidepressants (1)–(3) Behavioural activation (1)–(4) IPT (1)–(2) Shortterm psychodynamic psychotherapy (1)–(2) GP care (3) GP care.Table 36: (Continued) CBT versus comparator CBT versus rational emotive behavioural therapy (1)–(3) 100% MDD (1) 100% MDD (2) 46% mood disorder.85 mg/kg) .

(11) Nefazodone (300–600 mg/day) (12) Desipramine (90 mg/day) (13) Antidepressants (no details) (14) Paroxetine (10–50 mg) (15) Nortiptyline (25 mg) (16) Desipramine (150-300mg) (17) TAU (18) Desipramine (90 mg/day) (1) 15 weeks (2)–(4) 16 weeks (5) 10 weeks (6) 6 weeks (1) 14 weeks (1) 16 weeks (2) 12 weeks (3) 20 sessions (1)–(2) 16 weeks (1) 16 weeks (1) 20 weeks (3) Up to 16 weeks (2) 8 or 16 weeks (2) 16 weeks (4) 16 weeks (3) 6 weeks Length of treatment (1) 16–20 weeks (2) Maximum of 20 weeks (3) 16 weeks (4) 14 weeks (5)–(7) 16 weeks (8)–(9) 8 weeks (10) 10 weeks (11) 12 weeks (12) 18 weeks (13) 16 weeks (14) 8 weeks (15) 12 weeks (16) 16 weeks (17) 6 weeks (18) 8–12 weeks High-intensity psychological interventions Continued 221 .

222 Table 36: (Continued) CBT versus comparator CBT versus rational emotive behavioural therapy (1) 6 months (1) Not reported (2) 12 months (3) 6 months (1) 18 months (2)–(4) Not reported CBT versus behavioural activation CBT versus IPT CBT versus short-term psychodynamic psychotherapy (1) 12 months (2) Not reported CBT (primary care) versus GP care (1) 5 months (2) Not reported (3) 12 months (1) 12 months (2)–(3) Not reported (4) 18 months (5) Not reported (6) 2 months CBT versus antidepressants High-intensity psychological interventions Follow-up (1)–(2) Not reported (3) 24 months (4) 6 months (5)–(6) Not reported (7) 18 months (8)–(9) 12 months (10)–(11) Not reported (12) 6 months (13)–(16) Not reported (17) 12 months (18) No follow-up .

Table 37: Summary study characteristics of cognitive behavioural therapy CBT for the elderly CBT antidepressants versus CBT 6 RCTs (731) (1) Blackburn1981 (2) Hautzinger (in-pats) (3) Hautzinger1996 (4) Keller2000 (5) Murphy1984 (6) Thompson2001 (1) LAIDLAW2008 (2) Thompson2001 (1) Thompson2001 2 RCTs (104) 1 RCTs (69) CBT versus antidepressants CBT antidepressants versus antidepressants Group CBT versus waitlist control 1 RCT (45) (1) WILKINSON2009 CBT antidepressants versus antidepressants No. trials (total partici-pants) 9 RCTs (850) Study IDs (1) Blackburn1981 (2) FAVA1998* (3) Hautzinger (in-pats) (4) Hautzinger1996 (5) Keller2000 (6) Miller1989 (7) Murphy1984 (8) Scott1997 (9) Thompson2001 (1) Unextractable (2)–(3) 113/63 (4) 445/65 (5) 52/74 (6) 47/69 (1) 29/73 (2) 67/67 (1) 67/67 N/% female (1) Unextractable (2) 24/60 (3)–(4) 113/63 (5) 445/65 (6) 34/74 (7) 52/74 (8) 32/67 (9) 47/67 (1) 28/62 Continued High-intensity psychological interventions 223 .

224 Table 37: (Continued) CBT for the elderly CBT antidepressants versus CBT (1) 43 (2)–(3) 40 (4) 43 (5) 33 (6) 67 (1) 76 (2) 67 (1) 67 (1) 74 CBT versus antidepressants CBT antidepressants versus antidepressants Group CBT versus waitlist control (1) 100% MDD (2)–(3) 80% MDD. 20% dysthymia (4)–(6) 100% MDD (1)–(2) 100% MDD (1) 100% MDD (1) Remission from depressive episode (1) Amitriptyline/ clomipramine (2)–(3) Amitriptyline (150 mg/ day) (4) Nefazodone (300–600 mg/day) (5) Nortriptyline (25 mg) (6) Desipramine (1) TAU – GP care (2) Desipramine (90 mg/day) (1) Desipramine (90 mg/day) (1) TAU CBT antidepressants versus antidepressants Mean age (1) 43 (2) 47 (3)–(4) 40 (5) 44 (6) 37 (7) 33 (8) 41 (9) 62 High-intensity psychological interventions Diagnosis (1) 100% MDD (2) Remission after previous treatment (3)–(4) 80% MDD. 20% dysthymia (5)–(9) 100% MDD Comparator (1) Amitriptyline/ clomipramine (2) Range of antidepressants (3)–(4) Amitriptyline (150 mg/day) (5) Nefazodone (300–600 mg/day) .

(6) TCAs (7) Nortriptyline (25 mg) (8) TAU (9) Desipramine (90 mg/day) (1) Maximum 20 weeks (2)–(3) 8 weeks (4)–(5) 12 weeks (6) 8–12 weeks (1) 18 weeks (2) 8–12 weeks (1) 8–12 weeks (1) 8 weeks Length of treatment (1) Maximum 20 weeks (2) 20 weeks (3) 8 weeks (4) 4 weeks (5) 12 weeks (6) 20 weeks (7) 12 weeks (8) 6 weeks (9) 8–12 weeks (1) Not reported (2)–(3) 12 months (4)–(6) Not reported (1) 6 months (2) Not reported (1) Not reported Follow-up (1) Not reported (2) 24 months (3)–(4) 12 months (5)–(7) Not reported (8) 12 months (9) Not reported (1) 12 months * Follow-up to Fava1994 (study in the previous guideline). High-intensity psychological interventions 225 .

226 Relapse prevention studies Group CBT versus TAU CBT versus antidepressants CBT antidepressants versus antidepressants 1 RCT (132) MBCT versus antidepressants MBCT versus comparator 1 RCT (187) 1 RCT (180) 1 RCT (123) 3 RCTs (288) (1) BOCKTING2005 (1) HOLLON2005 (1) PERLIS2002 (1) KUYKEN 2008 (1) CRANE2008 (2) MA2004 (3) Teasdale2000 (1) 137/73 (1) Not reported (1) 72/55 (1) 94/76 (1) Not reported (2) 57/76 (3) 110/76 (1) 44 (1) Not reported (1) 40 (1) 49 (1) 45 (2) 44 (3) 43 Table 38: Summary study characteristics of cognitive behavioural therapies Group CBT versus other group therapies Group CBT versus waitlist High-intensity psychological interventions No. trials (total participants) 3 RCTs (144) 5 RCTs (451) Study IDs (1) Beutler1991 (2) Bright1999 (3) Covi1987 (1) ALLARTVANDAM2003 (2) Brown1984 (3) DALGARD2006 (4) HARINGSMA2006A (5) WONG2008 N/% female (1) 40/63 (2) 70/71 (3) 42/60 (1) 65/57 (2) 44/55 (3) 118/176 (4) 76/55 (5) 75/78 Mean age (1) 47 (2) 46 (3) 44 (1) 48 (2) 37 (3) 47 (4) 64 (5) 37 .

44% intermittent depressive disorder (RDC). 11% minor depressive disorder (RDC) (3) 100% unipolar depression (4) 60% MDD. 95% no diagnosis but BDI 10 (2) 44% MDD (RDC). 40% MDD anxiety (5) 100% depression (1) TAU (1) Paroxetine (mean 38 mg/day) (1) Fluoxetine (40 mg/day) clinical management (1) Antidepressants (no details) (1) Waitlist (2)–(3) TAU (1) Remission from (1) Responded to (1) Remission depression 10 weeks previous treatment from MDD (1) Remission from MDD (1) Remission from depression (2)–(3) Remission from depression (major) Comparator (1) Focus expressive psychotherapy (2) Mutual support group therapy (3) Group psychotherapy (1) 8 weeks (1) 12 months (1) 28 weeks (1) TAU – free to seek (2) Waitlist (3) TAU (4)–(5) Waitlist Length of treatment (1) 24 months (1) 24 months (1) 20 weeks (2) 10 weeks (3) 14 weeks (1) 12 weeks (2)–(3) 8 weeks (4)–(5) 10 weeks (1) 8 weeks (1)–(3) 8 weeks Follow-up (1) 3 months (2)–(3) Not reported (1) 12 months (2)–(3) 6 months (4)–(5) Not reported (1) Not reported (1) 15 months (1) 2–3 months (2)–(3) 12 months High-intensity psychological interventions 227 .Diagnosis (1) 100% MDD (2) 100% MDD or dysthymia (3) 100% MDD (1) 5% dysthymia.

01 to 0.17 to 1.06 ( 0.24 to 0. 193 K n 1.30 to 0.26 to 0.21) CBT 84.95) Low Low SMD 0. 113 K n 2.12) SMD 0. 405 CBT 81.85) CBT 37.97 to 2.18 At 8 weeks: SMD 0.02 SMD 0.45 to 0.37 to 0. participants K n 13.19 SMD 0 ( 0. 1480 – Forest plot number CBT 15.228 CBT versus placebo clinical management RR 0.01 ( 0.17 Low High Moderate Moderate Table 39: Summary evidence profile for cognitive behavioural therapies CBT versus comparator High-intensity psychological interventions CBT versus antidepressants CBT versus waitlist control Leaving study early for any reason RR 0.03 SMD 0.44 (0.37) 2.46 (0.29 (0.23) High K n 1. 108 Low Moderate RR 1.61) RR 1.56 (0.54 (0. 208 CBT 13. 66 CBT 07.33) Quality High High .41) CBT versus rational emotive behaviourral therapy CBT versus behavioural activation CBT versus IPT CBT versus CBT (primary short-term care) versus psychodynamic GP care psychotherapy RR 0.85) Moderate K n 3.63 to 0.05 – Depression self-report measures at endpoint SMD 0.04 SMD 0.75 (0.89 ( 1.40 to 3.21 (0.34 ( 0.24 to 1.91) Not reported Quality High – Number of studies.22 (0.91 to 1.46) High K n 3.33) RR 1.51 to 0.61) Moderate K n CBT 03.35 ( 1.12 to 1.83 to 0.15 ( 0.77) RR 0.

74 to 0.01 High-intensity psychological interventions 229 .13 ( 0.17 CBT 82.68 to 0.40 to 0.02 SMD 0.05 clinician-report ( 0. 92 CBT 14. 24 CBT 04.04 Number of studies.02 CBT 82.23 to 0. n 383 K 1.03 ( 0.03 CBT 14. n 43 K 3.04) SMD 0.45 (0.08) Quality High Low High K n – 2.32 ( 0. 121 K n Low HRSD 6: RR 0. participants CBT 04. 430 Low – – SMD 0.62 to 0.19 CBT 74.05 CBT 01. n 54 K 1.Number of studies.32) Not reported CBT 86. n 120 Forest plot number CBT 17. participants K n 13.06 to 0.19 K n 1.18 K n 1. n 121 K 1. 113 CBT 74.91) SMD 0.06 to 0. n 113 K 1.02 CBT 86.18 CBT 08.03 ( 0.33 ( 0.04 K 8.04 Depression SMD 0. 43 K n 4. n 57 K 2.57) Low 1.05 CBT 02. n 480 K 2.15) measures at endpoint Low 1.34) At 8 weeks: SMD 0. 1403 K n Forest plot number CBT 17.

05 ( 0. n Forest plot number CBT 33.79 to 0. n 69 Moderate Low Low K 1.21) RR 0.12 At 6 months: BDI 12: RR 1. participants K 6.31 ( 0.36 ( 0.38 ( 0.09 SMD 0.22) CBT 57.12 At 6 months: MADRS 10: RR 0.12 SMD 0.26 to 2. n 108 Moderate Low K 1.84 (0.Table 40: Summary evidence profile for cognitive behavioural therapies CBT for the elderly CBT antidepressants versus CBT RR 1.14) Quality High Number of studies.48 to 1.61 to 0.30) Moderate 831 CBT 45.93 to 0.12) K 5.31) Quality High Number of studies.65 to 1. n High-intensity psychological interventions Forest plot number CBT 31.07 Depression clinicianreport measures at endpoint SMD 0.17 ( 0. n CBT antidepressants versus antidepressants Leaving study early for any reason RR 0.26 (0.44 to 0.69 to 0.01) Quality Moderate Number of studies.41 ( 0.00 (0.14) High 220 K 2.03) 69 SMD 0.12 CBT 60. n CBT 61.12 SMD 0. n CBT 62.09 SMD 0.12 37 37 45 RR 0.92 (0.12 SMD 0. n CBT 62. n 108 K 1.31 to 0.46 ( 0. n CBT 61.10) Moderate 277 CBT 47.07 . 724 Forest plot number CBT 34. n 710 K 2.72) 230 Moderate K CBT 47.27 to 1.81 (0.21) Low K 1.07 Depression selfreport measures at endpoint SMD 0.03) K 4.84 to 0.03 to 2. n CBT 57.77 to 1.07) CBT 56.68 to 4.12 69 Low K 1. 219 K 2.57 (0.62 to 0.45 ( 0.69 (0.12 108 Low K 1.75) CBT versus antidepressants CBT antidepressants versus antidepressants Group CBT antidepressants versus antidepressants RR 0. participants K 8. participants K 7.09 4. n CBT 62.

98 to 0.22 – – CBT 102.35) High K 4.57 to 1.07 to 1.05) RR 1.20 (0. n 369 K 1.21 – – K 2. participants CBT 89.44 to 4.Table 41: Summary evidence profile for cognitive behavioural therapies Relapse prevention studies Group CBT versus waitlist Group CBT versus TAU CBT versus antidepressants CBT antidepressants versus antidepressants RR 0. n 42 Number of studies.61 to 0. n 180 K 1.94 (0. n 187 K Moderate Low RR 2. n 83 Forest plot number CBT 97.84 to 0.14 CBT 68.85) Leaving study early for any reason RR 0. n 132 Low K 1.15 Not reported K 3.16 CBT 65.52) Low 1.23 High-intensity psychological interventions Continued 231 .61 to 1.61) MBCT versus antidepressants MBCT versus comparator Group CBT versus other group therapies RR 1. n 158 Forest plot number CBT 96.21 SMD 0.11) Low K 4. participants CBT 90.25) Depression selfreport measures at endpoint SMD 0.96 (0.34 (0.17 ( 0.60 ( 0.53) Not reported Quality Moderate – – Number of studies.34 (0.26) Quality Moderate – – – – Low K 1.01 to 6.36 ( 0.47 (1.24 Not reported – At 1 month: SMD 0.30 to 4. n 123 RR 0. n 277 – – – – Not reported Not reported CBT 70.22 CBT 103.

participants K 2. n 123 Moderate K 1.39) Low K 1.55 to 0.11) Relapse: patients with 3 episodes: RR 0.18 ( 0.13 High-intensity psychological interventions Group CBT versus other group therapies Depression clinician-report measures at endpoint SMD 0.12 ( 0. n 83 Forest plot number CBT 97.46 (0. n 71 – – Low K 1.54 to 0.08 ( 0.15 CBT 66. n 132 Not reported Relapse RR 0.22 .14 CBT 64.52 to 0.79) Low K 1.16) MBCT versus antidepressants MBCT versus comparator Not reported Patients with 5 previous episodes: SMD 0.16 – CBT 69.31) Quality Moderate Number of studies.232 Table 41: (Continued) Relapse prevention studies Group CBT versus waitlist Group CBT versus TAU CBT versus antidepressants CBT antidepressants versus antidepressants SMD 0.27 to 0.80 (0.57 to 1. n 55 – – – CBT 72.

95% CI 0. 95% CI 1. suggests broad equivalence between CBT and antidepressants. along with the relatively narrow CIs. 0.65.06 to 0. A 1-year follow-up of the DIMIDJIAN2006 trial indicates that people who had cognitive therapy were less likely to relapse following treatment than those previously treated with antidepressants (RR 0. The results for depression scores at post-treatment (BDI: SMD 0. The combination treatment of CBT and antidepressants had a lower risk of discontinuation compared with antidepressants (RR 0. n 137) indicates that CBT has a significant medium effect (BDI: 0.24 to 0. In terms of leaving the study early.31) depression scores. However. there was limited data (BDI at 6 months: SMD 0.41.45.46.4 Clinical evidence summary for cognitive behavioural therapies Cognitive behavioural therapies versus antidepressants There were sixteen trials (n 1793) that reported the effectiveness of CBT compared with antidepressants. Combination (cognitive behavioural therapy antidepressants) versus antidepressants Nine studies included a comparison between combined treatment of CBT plus antidepressants and antidepressants alone.07).82.06. 0. 0.53.08.75.81. 95% CI 0.76. 0. 0.High-intensity psychological interventions 8. 95% CI 0.64.14) and clinician-rated (SMD 0.50.53. 95% CI 0. 95% CI 0. HRSD: SMD 0. 95% CI 0. HRSD at 12 months: SMD 0.29. BDI at 12 months: SMD 0. Cognitive behavioural therapies versus comparator (waitlist control) Four low quality studies (two reported in the previous guideline: Beach1992 and Selmi1990.50.01).33) in self-reports and showed an effect in clinician-reported depression scores (RR 0.62. there was a significant higher risk of discontinuation (RR 0. 95% CI 0.61. 0.15) over antidepressants. 1. 1.29.38.60. HRSD: SMD 0.70.45. There is evidence that the combined treatment has a significant medium effect in the reduction of self-rated (SMD 0. At 6 and 12 months’ follow-up. and two found in the update search: DERUBEIS2005 and DIMIDJIAN2006) compared the efficacy of cognitive behavioural therapies versus waitlist control.1.68 to 0. Six of those studies were found in the search of the guideline update and ten were reported in the previous guideline.11). 233 .65.40. 1. however.59 to 0.12. The effectiveness of CBT for the treatment of depression was large (SMD 0. 95% CI 0. 95% CI 0. 95% CI 0. by 12 months’ follow-up the evidence from three trials (Hautzinger [in-pats].91).07. HRSD at 6 months: SMD 0. 95% CI 0. 95% CI 0. 95% CI 0.69. 1.74) were not significantly different and this. 0.05. 0. CI 0.35. 95% CI 0.23.15) and at 1-month follow-up (BDI: SMD 0. Only one of those studies (FAVA199856) was found in the search for this guideline update. HRSD: 0.91) in the antidepressant group.12.89.84.63. which introduced some uncertainty 56Follow-up to Fava1994 (study in the previous guideline). Hautzinger1996 and Blackburn1997. 0. 95% CI 0.02.

There was some indication of higher dropout rates in the placebo groups but the effect (RR 0. This might suggest that although the CBT and antidepressants dataset supports combined treatment. 95% CI. 0. 95% CI 0.57). 95% CI 0. However. 1.01.82.60. Combination (cognitive behavioural therapy antidepressants) versus cognitive behavioural therapy Six studies reported in the previous guideline included a comparison of combination treatment and CBT alone. 95% CI 0.11).13. 0. there were no clinically important differences between CBT and IPT (BDI at endpoint: 0.32. There were no clinically important differences identified between CBT and behavioural activation (BDI at endpoint: 0. Cognitive behavioural therapies versus other therapies designed for depression (behavioural activation and interpersonal therapy) There were three studies that compared cognitive behavioural therapies with behavioural activation in the treatment of depression (DIMIDJIAN2006. 95% CI 0.29.06.44. 0. MARSHALL2008). 0. 0. In contrast with the dataset on the combination of CBT and antidepressants versus antidepressants. 0. JACOBSON1996). HRSD at endpoint: 0.61) was not significant and therefore inconclusive. BDI at 1-month follow-up: SMD 0.21 and clinician-rated: SMD 0. There was a small effect on reducing depression scores at endpoint in favour of CBT (self-rated: SMD 0. However.68. 0.34. HRSD at 1-month follow-up: SMD 0. HRSD at endpoint: 0.95.10. 95% CI 0.High-intensity psychological interventions about the relative long-term effectiveness of the combination of these two treatments. This evidence although limited suggests that IPT might be as effective as CBT in the treatment of depression.03. Gallagher1982.08. Again. A 1-year follow-up of the DIMIDJIAN2006 trial indicates that people who had cognitive therapy were less likely to relapse following treatment than those previously treated with antidepressants (RR 0. 95% CI 0. 95% CI 0. 95% CI 0. 0. the Gallagher1982 study only reported leaving study early data. 234 . the comparison in the Gallagher1982 study included cognitive therapy following the approach of Beck and colleagues (1979) and Emery (1981) and was compared with behavioural therapy that followed Lewinsohn’s (1975) approach.15.26. n 193). 1.32). it was not possible to identify a benefit for adding antidepressants to CBT (BDI at post-treatment: SMD 0.62.21.57). LUTY2007. 0. Freeman2002. 95% CI 0.12. From this evidence it is not possible to draw any clear conclusions about the relative efficacy of the treatments.72.94.51. Four studies included a comparison of CBT versus IPT (Elkin1989.04) when compared with placebo plus clinical management. the results were not significant and the CIs were fairly wide so the evidence remains inconclusive.36.17. In addition. Cognitive behavioural therapies versus comparator (placebo plus clinical management) There was little evidence of the increased effectiveness of CBT when compared with placebo plus clinical management from two studies (also reported in the previous guideline: Elkin1989 and Jarrett1999.41. 0.44. 95% CI 0. clinical benefit could still be derived from CBT alone.

30.33. Four studies evaluated the Coping with Depression programme (see above) (ALLARTVANDAM2003.34. However. 95% CI 0.12. Cognitive behavioural therapies (primary care) versus GP care Three trials reported in the previous guideline included a comparison between CBT in primary care versus usual GP care. The results show no significant difference in risk for discontinuation (RR 0. 1.83. HRSD: SMD 0. Two studies (Gallagher-Th1994 and Shapiro1994) compared CBT with short-term psychodynamic psychotherapy. At the end of treatment self-report depression scores (SMD 0.35. 0.85) were not significantly different. 95% CI 1. 95% CI 0. The evidence indicates no clinically important differences for the comparison of CBT with shortterm psychodynamic psychotherapy in decreasing depression (BDI at endpoint: SMD 0. In terms of leaving the study early due to any reason. 95% CI 0. BROWN1984. HARINGSMA2006). 95% CI 0. Klein1984) and no new studies were found that looked at this comparison for the guideline update58. The evidence indicates no clinically important difference in risk for discontinuation (RR 1. 235 .60. Group cognitive behavioural therapies Three studies reported in the previous guideline looked at the effectiveness of group CBT when compared with other psychotherapies (Bright1999. There was a significant medium effect of group CBT in lowering depression scores at endpoint (SMD 0.46. 95% CI 0.17. A further analysis was carried out looking at group CBT compared with waitlist control or treatment as usual.61. 0. 57For reasons of brevity this analysis is not included in the summary evidence table.79).26. 0.56. 0. 4. 0. 95% CI 0. DALGARD2006. 2.91).60.31).08).97.01. 95% CI 0.74.46). there was a significant difference favouring group CBT (RR 0. The evidence here is difficult to interpret as many patients in GP care might have been in receipt of antidepressants and the duration of treatment was shorter than that typical of CBT. 0. 0.57. The study characteristics are in Appendix 17b. The studies varied in duration: Freeman2002 consisted of 16 sessions over a 5-month period. 95% CI 0.44. the evidence suggests that there is a higher risk for discontinuation in those in the CBT (primary care) group (RR 1. Covi1987. Scott1992 was 16 weeks’ duration and Scott1997 was 6 weeks.11).53) or depression scores at post-treatment (BDI: SMD 0.94. 95% CI 0.61) or with Gestalt psychotherapy (BDI at endpoint: SMD 0.17.54. 58Ibid. One study (Rosner1999) compared CBT with gestalt psychotherapy57. and neither were clinician-rated depression scores (SMD 0. when self-rated depression scores were analysed by a cut-off of BDI 9. but can be found in Appendix 16b and 19b. From this evidence it is not possible to draw any clear conclusions about the relative efficacy of the treatments.High-intensity psychological interventions Cognitive behavioural therapies versus other psychotherapies not specifically designed for depression There were three studies that looked at the effectiveness of CBT compared with other therapies not specifically designed for depression.55.

A number of studies have addressed the issue of relapse prevention and have developed a number of different approaches both to the patient population identified and the specific CBT approach taken. 95% CI. 0.45.14) and this prevents any clear conclusion being drawn. The results suggest the effectiveness of CBT seen in adults of working age may be replicated in older adults but some caution is required in interpreting the results. 95% CI 0. individual CBT.84. 2.26. Thompson2001 also included a comparison of the combination of CBT with antidepressants with antidepressants alone.78. Montgomery–Åsberg Depression Rating Scale [MADRS] 10 at 6 months: RR 0. combination of CBT and antidepressants. Cognitive behavioural therapies – relapse prevention This section brings together the impact of relapse prevention studies in different areas (group CBT. 95% CI 1. Therefore group CBT (in particular Coping with Depression) appears an effective treatment for people with mild depression. the results were not significant for follow-up data (at 3 months: SMD 0. 6. 0. and MBCT). There is insufficient evidence (one study and wide CIs) to determine the 236 .79. 0. The evidence was inconclusive regarding leaving the study early.35. There were medium effects favouring combined treatments for both self-rated (SMD 0.93 to 0. In clinicianrated depression scores.03. In total. 0. 95% CI 0. 0. 95% CI 0.01. seven studies (n 957) found in the search for the guideline update examined relapse prevention in people who had been administered CBT.05). specific group-based approaches including a programme for those with residual symptoms (BOCKTING2005) and MBCT. 4. It should be noted that the CIs for both effects just cross the line of no effect. In the combined treatment of CBT plus antidepressants versus antidepressants alone. The approaches include extending the duration of individual CBT. n 43) comparing group CBT plus antidepressants with antidepressants alone in the treatment of depression for the elderly is not significant (BDI 12 at 6 months: RR 1.69.03). 95% CI 0.40. Three of these studies compare CBT with antidepressants. WILKINSON2009 looked at the effectiveness of group CBT in relapse prevention compared with waitlist control.High-intensity psychological interventions 95% CI 0. 95% CI 0.07 and at 6 months: 0.84 to 0. there was a significant medium effect favouring CBT (SMD 0.68.03). The evidence from one trial (WILKINSON2009.15.35) and at 6 months’ follow-up (SMD 0.47. LAIDLAW2008 and Thompson2001 compared CBT with antidepressants.74.92).41. 95% CI 0.02).36. there was little to no difference in risk for discontinuation between the two groups (RR 0. 95% CI.44). Group cognitive behavioural therapy versus treatment as usual The evidence from one study (BOCKTING2005) indicates a higher risk for discontinuation in those administered group CBT than treatment as usual (RR 2.12) and clinician-rated depression scores (SMD 0. However. 0. 0. Cognitive behavioural therapies for the elderly Three studies looked at the effectiveness of CBT in the treatment of depression in elderly populations.21. so these results should be interpreted with caution.83.

HRSD: SMD 0. MA2004. 95% CI 0. Regarding the reduction of relapse rates.94. 95% CI 0.80. 0. KUYKEN2008.79). 2008. PAYKEL2005.. the evidence indicates a non-significant difference in self-reports of depression in patients with five or more previous episodes of depression (SMD 0.01.. group MBCT when compared with antidepressants showed a small to medium effect of group MBCT lowering depression scores at 1-month (BDI: SMD 0.59. 8.02. 0.30. They report a significant difference favouring individual CBT in relapse rates when compared with clinical management (RR 0.23. data at 68 weeks should be interpreted with caution given that there is only one study and the CIs are wide.35) but a significantly lower risk for relapse (RR 0. It is important to mention that the study that reports this comparison is based on a series of post-hoc analyses and results should be interpreted with caution. 0. however. The evidence indicates a higher risk for discontinuation in the combined treatment (RR 19. 0. 0. These studies.61. Mindfulness-based cognitive therapy Four studies (CRANE2008. However.31. Similarly. Two studies (MA2004. 95% CI 0. 95% CI 2. 1. 95% CI 0.05) and at 15 months’ follow-up (BDI: SMD 0.37.12). HRSD: SMD 0. 95% CI 0.37. 95% CI 0. 95% CI 0.74. 1. Teasdale2000) compared the combined treatment of group MBCT with GP care versus GP care alone.11. 0. 0.69. 2.96).52) or relapse (RR 0. 141.96. 95% CI 0. Cognitive behavioural therapy versus clinical management (not shown in tables) Two studies (FAVA1998 and PAYKEL2005) compared the effectiveness of individual CBT with clinical management (with antidepressants) as part of a relapse prevention programme.72. reports remission at 68 weeks (RR 1. appear in the study characteristics tables in Appendix 17b and the forest plots in Appendix 19b.1. Combination cognitive behavioural therapy antidepressants versus antidepressants When the combination treatment of CBT plus antidepressants was compared with antidepressants alone there were no significant differences in terms of risk for discontinuation (RR 0.54. 95% CI 0.39).66.22.85).57. The two studies mentioned previously are not shown in the tables of study characteristics or in the summary of evidence profiles in this chapter in the interest of brevity and given that these studies report different outcomes from those listed in the tables.34.58. 95% CI 0.80). Details on the methods used for the systematic search 237 .High-intensity psychological interventions comparative effectiveness between the two groups in terms of relapse or remission rates at 68 weeks.08.5 Health economic evidence and considerations Two studies were identified in the systematic literature review that evaluated the cost effectiveness of cognitive behavioural therapies for people with depression (Kuyken et al.54. Scott et al. Teasdale2000) evaluated the effectiveness of MBCT group treatment in relapse prevention. 2003). Furthermore. one of the two studies.

group and marital therapy and medication for this period. Kuyken and colleagues (2008) evaluated the cost effectiveness of MBCT compared with maintenance antidepressant medication in 123 patients with a history of depression participating in a primary care-based RCT. with a mean difference of £779 per person (p 0.370 versus $2.328 per relapse averted or £12. 8. Evidence tables for all health economics studies are presented in Appendix 15. Costs included all hospital care. Scott and colleagues (2003) evaluated the cost effectiveness of cognitive therapy added to antidepressants and clinical management compared with antidepressants and clinical management alone in a UK RCT of 154 patients with partially remitted major depression. the ICER was $962 per relapse/recurrence prevented and $50 per depression-free day. depression is seen as the result of a low rate of positive reinforcement and is maintained through negative reinforcement. hospital care. the cognitive therapy group was significantly more costly than standard clinical treatment. From a societal perspective. In behavioural therapies. 1979) there have been no treatment trials of this approach. The outcome measures used in the cost-effectiveness analysis were the mean total number of relapses/recurrences avoided and the mean total number of depression-free days. 1971) with treatment recommendations (Wolpe. community health and social services and any productivity losses resulting from time off work. primary care. The authors suggested that the additional cost of MBCT may be justified in terms of improvements in the proportion of patients who relapsed.6. Over 15 months’ follow-up. The time horizon of the analysis was 15 months and both a health service and societal perspective were taken in separate analyses.2.2 8. Operant or instrumental learning posits that depressive behaviours are learned through the contingencies around those behaviours.915. Although classical conditioning theories for depression have been put forward (for example. The ICER of cognitive therapy versus standard care was £4.865). there was no significant difference in total mean costs between MBCT and antidepressant treatment (US $3. The primary outcomes used in the analysis were relapse rates for the two treatment groups. adjunctive cognitive therapy was more costly but more effective than intensive clinical treatment alone.50 per additional relapse-free day. From an NHS and Personal Social Services (PSS) perspective. Section 3. The authors concluded that in individuals with depressive symptoms that are resistant to standard treatment. Wolpe. The setting was either in local clinics or in participants’ homes. p 0. Most commonly. clinical management. 238 .01). Overall. the ICER was $429 per relapse/recurrence prevented and $23 per depression-free day.1. The study estimated NHS costs including treatments. The time horizon of the analysis was 68 weeks (including 20 weeks of treatment).High-intensity psychological interventions of the health economics literature are described in Chapter 3.1 BEHAVIOURAL ACTIVATION Introduction Behavioural activation for depression evolved from learning theory that posits two types of learning: operant or instrumental learning and classical conditioning.

Hopko et al. four were found in the searches for the guideline update and two were from the previous guideline. graded exposure. entailed a comparison between behavioural activation and antidepressants.1). References for studies from the previous guideline are in Appendix 18.. 8. as a therapy in its own right.High-intensity psychological interventions patients use avoidance to minimise negative emotions and situations they worry will be unpleasant. feelings states and/or problem areas ● seek to reduce symptoms and problematic behaviours through behavioural tasks related to: reducing avoidance. which were excluded: CULLEN2006 because of a lack of extractable data60 and Thompson 1987 because it was unclear which patient numbers were used in their table reporting outcome measures and the dropout data was not fully reported. Comparisons between behavioural activation and cognitive behavioural therapies can be found in the previous section (see Section 8. Two further studies were identified. with full details in Appendix 17b..2. Behavioural therapies focus on behavioural activation aimed at encouraging the patient to develop more rewarding and task-focused behaviours as well as stepping out of patterns of negative reinforcement. 60The review team contacted the authors of the study but did not receive the data. activity scheduling.. Summary study characteristics of the included studies are presented in Table 42.1 and therefore have not been included in Table 42. entailed a comparison with psychotherapy. Studies comparing CBT and behavioural activation are reported in Section 8. DIMIDJIAN2006. derived from the behavioural model of affective disorders and where the therapist and patient: ● work collaboratively to identify the effects of behaviours on current symptoms. The approach was developed by Lewinsohn (1975). Jacobson et al. structured psychological intervention. as it is now known. 239 . McLean1979. as well as a comparison between behavioural activation and placebo. HOPKO2003 compared behavioural activation with an attentional control (the control had the same duration of contact in a group but no therapy was given) in an inpatient setting. One study. time-limited. 2001. 1997). In recent years there has been renewed interest in behavioural activation (for example. and initiating positively reinforced behaviours. 2003).2 Studies considered59 There were six studies involving a comparison of behavioural activation. Definition Behavioural activation is defined as a discrete. A further study. 59Study IDs in title case refer to studies included in the previous guideline and study IDs in capital letters refer to studies found and included in this guideline update. which also includes details of excluded studies. Of these. although it has always been part of cognitive behavioural treatments of depression (Beck et al.

respectively.High-intensity psychological interventions Table 42: Summary study characteristics of behavioural activation Behavioural activation versus placebo No.17 mg/day paroxetine (1) DIMIDJIAN2006 (1) HOPKO2003 (2) McLean1979 (1) 159/65 (1) 40 (1) 100% major depression (1) Placebo (1) 25/36 (2) 111/72 (1) 30 (2) 39 (1)–(2) 100% major depression (1) Supportive psychotherapy (2) Psychodynamic psychotherapy (1) 2 weeks (2) 10 weeks (1) None reported (2) 3 months Length of treatment Follow-up (1) 16 weeks (1) None reported (1) 16 weeks (1) None reported 8. 8. The full evidence profiles and associated forest plots can be found in Appendix 16b and Appendix 19b. Gallagher1982 only reported 61Cognitive therapy based on Beck and colleagues (1979) and Emery (1981) and compared with behaviour therapy based on Lewinsohn (1975). 240 .4 Clinical evidence summary Behavioural activation versus cognitive behavioural therapy Studies comparing behavioural activation and CBT are reported in the CBT summary evidence profile tables – see the section ‘Cognitive behavioural therapies versus other therapies designed for depression (behavioural activation and interpersonal therapy)’. In summary.2. there were three studies included (DIMIDJIAN2006. Gallagher198261 and JACOBSON1996). trials (total participants) Study IDs N/% female Mean age Diagnosis Comparator 1 RCT (96) Behavioural activation versus comparator 2 RCTs (136) Behavioural activation versus antidepressants 1 RCT (159) (1) DIMIDJIAN2006 (1) 159/65 (1) 40 (1) 100% major depression (1) 35.2.3 Clinical evidence Because there are a relatively small number of studies for behavioural activation a summary of evidence profile table has not been included here.

95% CI 0. Similarly. BDI: SMD 0. HRSD: SMD 0.78 and clinician-reported scores: SMD 0. McLean1979. HOPKO2003. 95% CI 0. 4. compared behavioural activation with a supportive treatment (three times weekly.76 and clinician-reported scores: SMD 0.14.62.High-intensity psychological interventions leaving the study early data. only leaving the study early data could be extracted. 0. 0. 0.73). Behavioural activation versus other interventions One study. 95% CI.14). 95% CI 0.34.17.07. 95% CI 0. There seems to be little to no difference between behavioural activation and antidepressants in terms of relapse rates at 1 year (RR 1.57). 0.64). 95% CI 0.49).12. the results were not significant but tended to favour the antidepressant group in those diagnosed with moderate severity (self-reported scores: SMD 0. The evidence suggests there is no significant difference between treatments in risk for discontinuation (RR 1. 0. From this study.23.31.04.21). In terms of depression scores. 241 . 95% CI. Behavioural activation versus placebo Only one study (DIMIDJIAN2006) included a comparison of behavioural activation versus placebo. The study used a short-term psychotherapy (10 weeks of 1-hour sessions) following Marmor (1973. 0.49. 0. 20 minutes for 14 days). 95% CI 0.49. 0. 2.83). 95% CI 0. 0. compared behavioural activation with an attentional control. 0. Behavioural activation versus antidepressants There is limited evidence from one study (DIMIDJIAM2006) of the effect of behavioural activation in the treatment of depression when compared with antidepressants. However. 1975) and Wolberg (1967). there were no significant differences between treatments in the reduction of depression scores (self-reported.56. There were no clinically important differences identified between CBT and behavioural activation (BDI at endpoint: 0.69. 95% CI. The second study. 95% CI.33. 1.04.26.95.71). 0. HRSD at endpoint: 0.77) and in those with high severity (self-reported scores: SMD 0.75 and clinician reported.52.62. The results at post-treatment favoured behavioural activation (BDI: SMD 0.47.61. These results are based on one mediumsized study and given its wide CIs it is difficult to make any firm conclusions from this evidence.24.15.04. this result is not significant and should be interpreted with caution. 95% CI 0. 0. the aim of which is the development of insight of the psychodynamic forces that initiated the patient’s current depression. 0. It should be noted that this evidence is based on one study and the CIs are wide.03. 95% CI 0. 0. and their results indicate an increased risk for discontinuation in the control group (RR 0. This limited evidence seems to indicate a low risk of discontinuation in the people administered antidepressants when compared with those in the behavioural activation group (RR 0.29. From this evidence it is not possible to draw any clear conclusions about the relative efficacy of the treatments.06. which was a non-directive discussion with the clinician in which the patient was encouraged to share their experiences.

relating to the effects of depressed state. As a consequence. 242 . Section 3. which also includes details of excluded studies. was excluded from this update because not all patients met criteria for depression ( 80%).3.1 PROBLEM SOLVING Introduction It has long been recognised that depression is associated with social problem-solving difficulties (Nezu. There has been recent interest in developing problem-solving therapies for depression for use in primary care (Barrett et al.. lack of knowledge. 8.3.2. References for studies from the previous guideline are in Appendix 18. The reasons for this may be various. 8. and rumination. Details on the methods used for the systematic search of the economic literature are described in Chapter 3. 1999. with full details in Appendix 17b. Summary study characteristics of the included studies are presented in Table 43. 62Study IDs in title case refer to studies included in the previous guideline and study IDs in capital letters refer to studies found and included in this guideline update. Mynors-Wallis2000).5 Health economic evidence and considerations No evidence on the cost effectiveness of behavioural activation for people with depression was identified by the systematic search of the economic literature..High-intensity psychological interventions 8. 1989). which focuses on learning to cope with specific problems areas and where therapist and patient work collaboratively to identify and prioritise key problem areas.. Dowrick et al. to break problems down into specific. Three studies were reported in the previous guideline but only two are included in the update (Mynors-Wallis1995. time-limited. Definition Problem-solving therapy is a discrete. 2000). helping patients solve problems and develop problem-solving skills has been a focus for therapeutic intervention and development of therapy (Nezu et al. structured psychological intervention. Two studies were found and excluded on the basis of one study not reporting the outcome data (AREAN2008) and one study having a sample size of less than ten (NEZU1986). and develop appropriate coping behaviours for problems.2 Studies considered62 No new studies found in the search for the guideline update were included. manageable tasks.3 8.6.1. problem solve. 1987). Dowrick and colleagues (2000) which was included in the previous guideline.

8.4 Clinical evidence summary Only two studies were found that met the inclusion criteria for problem solving and only one study (Mynors-Wallis1995) indicated that this intervention had a significant 243 . trials (total participants) Study IDs 1 RCT (70) (1) MynorsWallis1995 N/% female Mean age Diagnosis (1) 70/77 (1) 37 (1) 100% RDC MDD (1) 116/77 (1) 35 (1) 100% depression (1) 116/77 (1) 35 (1) 100% depression Comparator (1) Placebo (1) Amitriptyline (1) Fluvoxamine/ (150 mg/day) paroxetine (2) Fluvoxamine/ paroxetine (1)–(2) 12 weeks (1) 12 weeks (1) 12 months (1) Problem solving delivered by a nurse (as opposed to GP) (1) 12 weeks (1) 12 months Length of treatment Follow-up (1) 12 weeks (1) Not reported (1) Not reported (2) 12 months 8.High-intensity psychological interventions Table 43: Summary study characteristics of problem solving Problem solving versus placebo Problem Problem solving versus solving antidepressants antidepressants versus antidepressants 2 RCTs (135) (1) MynorsWallis1995 (2) MynorsWallis2000 (1) 70/77 (2) 116/77 (1) 37 (2) 35 (1) 100% RDC MDD (2) 100% depression 1 RCT (74) (1) MynorsWallis2000 Problem solving (GP) versus problem solving (nurse) 1 RCT (80) (1) MynorsWallis2000 No. respectively.3. The full evidence profiles and associated forest plots can be found in Appendix 16b and Appendix 19b.3 Clinical evidence Evidence from the important outcomes and overall quality of evidence are presented in Table 44.3.

73 to 0.41 to 1.21.24 ( 0.64 (0.69 ( 1. n 124 PS 19.33 to 0. n 60 PS 08.01 SMD 0.14) BDI 8: RR 0.54 to 0.High-intensity psychological interventions Table 44: Summary evidence profile for problem solving Problem solving versus placebo Problem solving Problem solving versus antidepantidepressants ressants versus antidepressants Problem solving (GP) versus problem solving (nurse) RR 1.67 (0.02 & PS 13. participants Forest plot number Depression clinicianreport measures at endpoint Quality Number of studies.65 to 2.09) Moderate Low K 2.03 (0.55 (0.39) Moderate Low K 2.18 to 4.44) HRSD 7: RR 1.05 (0.24.03 SMD 0.07 ( 0.67) Low Low K 1.66 ( 1.04 244 .01 & PS 05. n 70 PS 04.02 PS 19.02 SMD 0.20) Low K 2.01 SMD 0. n 71 K 1.80 to 3.66 to 1.02 ( 0.03 & PS 18. n 124 PS 16.25) BDI 8: RR 0.67) HRSD 7: RR 1. n 60 RR 0.18 ( 0. n 80 Leaving study RR 0.49 to 0.44) any reason Quality Number of studies.99) Moderate Moderate K 1.03 SMD 0.30 to 0.11 early for (0.03 to 0.02 & PS 12. n 71 PS 26.04 SMD 0.22) Low Low K 1.37 to 2. participants Forest plot number Moderate K 1. 0.39 to 0.04 SMD 0. n 65 Low K 1.04 & PS 24. participants Forest plot number Depression self-report measures at endpoint Quality Number of studies.45) HRSD 7: RR 1.01 & PS 06.89) Low K PS 01.89) Moderate Moderate K 1.11 ( 0. n 60 PS 11.01 PS 10.20 (0.62 (0.46 to 0.43 (0.34) Low 1.12) HRSD 7: RR 0. n 80 PS 03. 0.24) PS 22.88 (0.03 PS 26.02 SMD 0.10 ( 0.40) Low K 1.25 to 0. n 177 RR 1.85 to 2.

0.21. The mixed population in this study limits its relevance to this guideline. There were no significant differences when problem solving was compared with antidepressants or when the combination treatment of problem solving and antidepressants was compared with antidepressants alone.5 Health economic evidence and considerations One study was identified in the systematic literature review that evaluated the cost effectiveness of problem solving for people with common mental health problems (including depression and anxiety disorders) (Kendrick et al.12. 95% CI 0.6. 95% CI 0. estimated using utility scores derived from the EQ-5D questionnaire.83.62. The outcome measures used in the analysis were QALYs. social worker and psychiatrist.14) when compared with placebo.39. 95% CI 0. 1.001).66. 245 . 0. Costs estimated in each treatment group included nurse training and supervision.24. The setting was primary care and the study population included adult patients with a new episode of anxiety. A further study (Dowrick2000) indicated a significant decrease in the number of people diagnosed with depressive and subthreshold depressive symptoms after 6 months of treatment (RR 0.69. the mean cost difference between the problem solving and GP groups was £315 per patient (p 0. 0.68. hospital care plus out-of-pocket patient costs and productivity losses due to time off work. This effect was also seen for dichotomous scores: clinician-rated (RR 0. Kendrick and colleagues (2006a) evaluated the cost effectiveness of problem solving delivered by community mental health nurses (CMHNs) compared with usual GP care and generic CMHN care. No significant differences in utility scores or QALYs were detected between the three treatment groups at 26 weeks’ follow-up. 95% CI 1.3.89) and self-rated (RR0. 95% CI 1. However this trial did not meet the inclusion criteria for the guideline due to 80% or more of the population in the trial not meeting diagnosis for depression and therefore does not appear in the tables above.High-intensity psychological interventions effect in reducing depression scores (clinician-rated: SMD –0. Section 3. primary care. 8. Overall. The time horizon of the analysis was 26 weeks and two separate analyses were undertaken from a health service and societal perspective. but the uncertainty surrounding these results makes it difficult to draw any conclusions. depression or reaction to life difficulties (33% with a primary diagnosis of depression).99). self-rated: SMD –0. Details on the methods used for the systematic search of the health economics literature are described in Chapter 3.55.02) when compared with placebo.1.33. The results of the incremental analysis showed that both problem solving and generic CMHN care were dominated by GP care. References to included studies and evidence tables for all health economics studies are presented in the form of evidence tables in Appendix 15. Total direct health service costs and productivity losses were higher over 26 weeks in the problem solving group compared with GP or CMHN care. 0. 2006a)..

Hayhurst et al. the focus of the search was not on a specific approach but on couples therapy more generally.1 COUPLES THERAPY Introduction Therapists have noted that a partner’s critical behaviour may trigger an episode of depression. Difficulties in developing intimacy. Hooley & Teasdale. psychological intervention derived from a model of the interactional processes in relationships where: ● the intervention aims to help participants understand the effects of their interactions on each other as factors in the development and/or maintenance of symptoms and problems ● the aim is to change the nature of the interactions so that the participants may develop more supportive and less conflictual relationships.4. was excluded from the update because more than 50% of the participants dropped out from one arm of the study. with full details in Appendix 17b. Systemic couples therapy aims to give the couple new perspectives on the presenting problem (for example. and coping with conflict.. with physical aggression by a partner predicting depression in women. 8. also predict depression in both men and women (Christian et al. 1989). and explore new ways of relating (Jones & Asen. 1999). There has also been some research looking at differences in the vulnerabilities between men and women within an intimate relationship. 63Study IDs in title case refer to studies included in the previous guideline and study IDs in capital letters refer to studies found and included in this guideline update. depressing behaviours).4 8.4. 246 . although other researchers have questioned this (for example. 2000). Other developments such as those by Jacobson and colleagues (1993) took a more behavioural approach. Definition Couples therapy is defined as a time-limited. One study (Leff et al. In the analysis of couples therapy in this guideline. this study used a systemic approach based on the Jones and Asen (1999) manual.. Summary study characteristics of the included studies are presented in Table 45. which also includes details of excluded studies. Like other therapies..2 Studies considered63 Six RCTs were included in the review of couples therapy. 1994). References for studies from the previous guideline are in Appendix 18.High-intensity psychological interventions 8. Two studies were found in the search for the guideline update (BODENMANN2008 and JACOBSON1993) and four were also included in the previous guideline. couples therapy has evolved in recent years. which was included in the previous guideline. 1997). and/or maintain or exacerbate relapse in the long term (for example.

9% dysthymia (2) 89% MDD. 11% dysthymia (1)–(4) CBT (1) 20 weeks (2) 16 sessions (3) 20 weeks (4) 16 weeks (1) 18 months (2) Not reported (3)–(4) 12 months (1) 45 (2) 40 (1) MDD or dysthymia (2) 100% RDC MDD (1)–(2) IPT (1) 20 weeks (2) 16 weeks (1) 39 (1) 35/58 (2) 13/72 (1) 60/100 (1) BODENMANN2008 (2) Foley1989 2 RCTs (109) No. 11% dysthymia (1) 100% MDD (1) 100% MDD Comparator (1)–(2) Waitlist control (1) CBT (1) 20 sessions (1) CBT (1) 20 sessions Length of treatment (1) 15 weeks (2) 16 weeks High-intensity psychological interventions 247 Follow-up (1)–(2) 12 months (2) 12 months (1) 18 months (2) Not reported (1) 12 months (1) 12 months .Table 45: Summary study characteristics of couples therapy Couples therapy versus CBT Couples therapy versus IPT Couples therapy CBT versus CBT 1 RCT (41) (1) JACOBSON 1993 1 RCT (40) (1) JACOBSON 1993 Couples therapy CBT versus couples therapy Couples therapy versus waitlist control 4 RCTs (130) (1) BODENMANN2008 (2) EmanuelsZuurveen 1996 (3) JACOBSON1993 (4) O’Leary1990 (1) 35/58 (2) 14/52 (3) 60/100 (4) 36/100 (1) 45 (2) 38 (3)–(4) 39 (1) MDD or dysthymia (2)–(3) 100% MDD (4) 89% MDD. trials (total participants) 2 RCTs (81) Study IDs (1) Beach1992 (2) O’Leary1990 N/% female (1) 45/100 (2) 36/100 (1) 60/100 Mean age (1)–(2) 39 (1) 39 Diagnosis (1) 91% MDD.

K participants Forest plot number CT O1.51 to 0.67 (0.68 to 0. – participants Forest plot number – Depression selfreport measures at endpoint Quality SMD 1.01 ( 0.07 ( 0.01 Depression clinician-report measures at endpoint Quality Not reported CT O3. n 101 Couples therapy versus IPT RR 0. – participants Forest plot number – 248 CT O3. n 58 – Number of studies.4.56) Low K 1.75) High 2.38) Moderate K 2. The full evidence profiles and associated forest plots can be found in Appendix 16b and Appendix 19b.05 SMD 0.02 SMD 0.4 Clinical evidence summary In five of the studies included in this review the model used was a behavioural model. 8. n 40 Number of studies.52) – Low K 1.55 to 2.High-intensity psychological interventions 8. n 40 Moderate K 2.06 ( 0. respectively.4.10 ( 0.58 to 0. two other studies used a model based on IPT.22 to 2.69 to 0.35 ( 1. n 67 CT 12.05 . n 54 CT O2.05 SMD 0. Two studies (Beach1992 and Table 46: Summary evidence profile for couples therapy Couples therapy versus waitlist control Leaving study early for any reason Quality Not reported Couples therapy versus CBT RR 1.22 (0.02 CT 13.71) Moderate K 3.95 to 0.04) Moderate K 2.02 SMD 0. n 58 Number of studies.3 Clinical evidence Evidence from the important outcomes and overall quality of evidence is presented in Table 46.55) CT 13.

IPT is appropriate when a person has a key area of difficulty that is specified by the treatment (for example. patient and therapist agree to work on a particular focal area that would include: interpersonal role transitions.55.High-intensity psychological interventions O’Leary1990) indicated a significant large effect in reducing depression self-report scores at post-treatment (SMD 1. 95% CI 0.05.57) suggesting couples therapy has broadly similar effects to CBT. Section 3.5 8.58. not past ones. 0.22.6.38. grief or interpersonal conflicts). 95% CI 1. 0.10. 95% CI 0.95..5. by appropriately addressing interpersonal situations. or unconscious conflicts as in psychodynamic psychotherapy). 0.54). 2.69.90 to 0. It can be delivered as an individually focused therapy but has also been developed as a group therapy (Wilfley et al. grief and/or interpersonal deficits. and on interpersonal processes rather than intra-psychic ones (such as negative core beliefs or automatic thoughts as in CBT.41. The main clinical tasks are to help patients to learn to link their mood with their interpersonal contacts and to recognise that. they may simultaneously improve both their relationships and their depressive state.08. In a larger dataset where couples therapy is compared with individual CBT.75) when compared with waitlist control.71) or depression scores at post-treatment (BDI: SMD 0. It is time limited and focused on difficulties arising in the daily experience of maintaining relationships and resolving difficulties during an episode of major depression. 0. 8. HRSD: 0.55) or at 6 months’ follow-up (BDI: SMD 0. Details on the methods used for the systematic search of the economic literature are described in Chapter 3. 95% CI 0. there were no significant differences in risk for discontinuation (RR 1. The results from these two small-sized studies had wide CIs and therefore it is difficult to interpret the comparison of the two treatments with any confidence. interpersonal roles/conflicts.1 INTERPERSONAL THERAPY Introduction Interpersonal therapy (IPT) was developed by Klerman and Weissman (Klerman et al.07.. 95% CI 0.70. 95% CI 0.. 95% CI 0. 2000).35.1.67.09) but this does not persist to 1 and a half years (SMD 0. 0.5 Health economic evidence and considerations No evidence on the cost effectiveness of couples therapy for people with depression was identified by the systematic search of the economic literature.4. Two studies (BODENMANN2008 and Foley1989) compared couples therapy with IPT. 2000). IPT focuses on current relationships. Early in the treatment. There is some indication of an effect in reducing selfreported depression scores at 1 year’s follow-up (SMD 0. 249 . 8. 1984) initially for depression although it has now been extended to other disorders (Weissman et al.

High-intensity psychological interventions The character of the therapy sessions is. Table 48. Table 49 and Table 50 with full details in Appendix 17b. The most common reasons for exclusion were: that the trials did not report the outcome data. Maintenance treatment occurs when the episode is considered to have remitted or significantly improved. then as a continuation treatment. time-limited. that they included populations without a diagnosis of depression and that they used an unclear control intervention. two studies looked at IPT as a 3year maintenance treatment. there is an emphasis on effecting changes in interpersonal relationships and tasks towards this end may be undertaken between sessions. Three studies included a comparison of IPT with CBT. six were found in the new search for the guideline update and eight were also included in the previous guideline. largely. the patient is stable. facilitating understanding of recent events in interpersonal terms and exploring alternative ways of handling interpersonal situations. In this guideline continuation treatment is defined as a treatment that occurs after the acute symptoms have subsided. and these results are reported in Section 8. Summary study characteristics of the included studies are presented in Table 47. and finally. structured psychological intervention. for those who recovered. role transitions. 14 were included and eight were excluded. and four studies looked at IPT in an older population. 8. Although there is not an explicit emphasis on ‘homework’. grief and loss. References for studies from the previous guideline are in Appendix 18. It is important to mention that one study. 250 . feelings states and/or problems ● seek to reduce symptoms by learning to cope with or resolve these interpersonal problem areas.2 Studies considered64 Twenty-two trials were identified. but treatment is continued to avoid recurrence.5. derived from the interpersonal model of affective disorders that focuses on interpersonal issues and where the therapist and patient: ● work collaboratively to identify the effects of key problematic areas related to interpersonal conflicts. Definition IPT was defined as a discrete. is a four-arm trial of an elderly population. Of the 14 studies that were included. which also includes details of excluded studies. when the patient could be considered to be substantially improved and the aim is to achieve remission or significant improvements in symptoms and restore normal function. 64Study IDs in title case refer to studies included in the previous guideline and study IDs in capital letters refer to studies found and included in this guideline update. and social skills. they were randomised to IPT as a maintenance treatment. Reynolds1999.1. including IPT as an acute treatment. From the 14 included studies there were three examining IPT as a continuation treatment. and their effects on current symptoms. The terms ‘continuation’ and ‘maintenance’ have been used interchangeably in many trials.

Table 47: Summary study characteristics of IPT IPT versus GP care (including antidepressants) 4 RCTs (391) (1) Freeman2002 (2) MARSHALL2008 (3) Schulberg1996 (4) SWARTZ2008 (1) 96/61 (2) 70/69 (3) 229/83 (4) 47/100 (1) 37 (2) Not reported (3) 38 (4) 42 (1)–(4) 100% MDD (1) 68 (2) 70 (1) 80/75 (2) 25/71 (1) 96/63 (2) 28/80 (3) 80/75 (4) 81/65 (1)–(2) 35 (3) 70 (4) 41 (1) 100% MDD (2) 100% dysthymia (3)–(4) 100% MDD (1) Reynolds1999 (2) Weissman1992 2 RCTs (78) 4 RCTs (299) (1) BLOM2007 (2) de Mello2001 (3) Reynolds1999 (4) SCHRAMM2007 IPT versus IPT antidepressants IPT antidepressants versus antidepressants IPT versus antidepressants IPT versus placebo No. trials (total participants) 1 RCT (123) 3 RCTs (347) (1) Elkin1989 (2) Reynolds1999 (3) Schulberg1996 (1) 168/70 (2) 80/75 (3) 229/83 (1) 35 (2) 70 (3) 38 (1) 100% RDC MDD (2)–(3) 100% MDD Study IDs (1) Elkin1989 N/% female (1) 168/70 Mean age (1) 35 Diagnosis (1) 100% RDC MDD (1) 100% MDD (2) 100% moderate/ severe MDD High-intensity psychological interventions Continued 251 .

252 Table 47: (Continued) IPT versus GP care (including antidepressants) (1)–(4) GP care (1) Nortriptyline (2) Alprazolam (2.5 mg/day) (1) 16 weeks (2) 6 weeks (1) 14 weeks (2) 6 months (3) 16 weeks (4) 5 weeks (1) Not reported (2) 5 months (3) Not repoerted (4) 12 months (1) Nefazodone (2) Moclobemide (150–300 mg/day) (3) Nortriptyline (4) Sertraline (90 mg/day) IPT versus IPT antidepressants IPT antidepressants versus antidepressants IPT versus antidepressants (1) Imipramine (150 mg/day) (2) Nortriptyline (3) Nortriptyline (190–270 mg/day) (1)–(2) 16 weeks (3) 4 months (1) 5 months (2) 16 weeks (3) 4 months (4) Not reported (1) 5 months (2) Not reported (3) 4 months (4) 9 months (1)–(2) Not reported (1)–(2) Not reported (3) 4 months IPT versus placebo High-intensity psychological interventions Comparator (1) Placebo Length of treatment (1) 16 weeks Follow-up (1) Not reported .2 mg/day) or imipramine (97.

trials (total partici-pants) (1) Schulberg1996 (1) 229/83 (1) 38 (1) 100% MDD (1) TAU (1) 4 months (1) 4 months 1 RCT (184) 1 RCT (185) Study IDs (1) Schulberg1996 N/% female (1) 229/83 Mean age (1) 38 Diagnosis (1) 100% MDD (1) 100% dysthymia (1) 100% MDD (1) Moclobemide (150–300 mg/day) (1) 6 months Comparator (1) Nortriptyline (190–270 mg/day) Length of treatment (1) 4 months Follow-up (1) 4 months High-intensity psychological interventions 253 .Table 48: Summary study characteristics of IPT as continuation treatment (up to 6 months) IPT as continuation treatment IPT versus TAU IPT antidepressants versus antidepressants 1 RCT (35) (1) de Mello2001 (1) 28/80 (1) 35 1 RCT (43) (1) Reynolds1999 (1) 80/75 (1) 70 IPT antidepressants versus IPT placebo IPT placebo versus medication clinic placebo 1 RCT (50) (1) Reynolds1999 (1) 80/75 (1) 70 (1) 100% MDD (1) Nortriptyline (1) 16 weeks (1) 5 months (1) Not reported (1) Placebo (1) 16 weeks (1) Not reported IPT versus antidepressants No.

14% bipolar (2) Recovered after continuation treatment (1) Imipramine (150–300 mg/day) (1) 3 years (1) Not reported (1) Placebo (1) 10 week remission from MDD (100%). 14% bipolar (2) Recovered after continuation treatment (1) Imipramine (150–300 mg/day) (2) Nortriptyline (1) 3 years (1) Not reported (1)–(2) 3 years (1)–(2) Not reported (1) 98/77 (1) 98/77 (1) 98/77 (2) 80/75 (1) Frank1990 (1) Frank1990 (1) Frank1990 (2) Reynolds 1999B (1) Frank1990 (2) Reynolds 1999B (1) 98/77 (2) 80/75 (1) 40 (2) 68 (1) 10 week remission from MDD (100%). 14% bipolar (1) 10 week remission from MDD (100%). trials (total participants) 1 RCT (51) Study IDs (1) Frank1990 High-intensity psychological interventions N/% female (1) 98/77 Mean age (1) 40 Diagnosis (1) 10 week remission from MDD (100%). 14% bipolar Comparator (1) Imipramine (1) Imipramine (150–300 mg/day) (150–300 mg/day) (2) Nortriptyline Length of treatment (1) 3 years Follow-up (1) Not reported . 14% bipolar (1) 40 (1) 40 (1) 40 (2) 68 (1) 10 week remission from MDD (100%).Table 49: Summary study characteristics of IPT as maintenance treatment (3 years) IPT as maintenance treatment (3 years) IPT antidepressants versus IPT placebo 2 RCTs (94) (1) Frank1990 (2) Reynolds 1999B (1) 98/77 (2) 80/75 (1) 40 (2) 68 (1) 10 week remission from MDD (100%). 14% bipolar (2) Recovered after continuation treatment (1) Imipramine (150–300 mg/day) (2) Nortriptyline (1)–(2) 3 years (1)–(2) Not reported 1 RCT (54) 1 RCT (49) 2 RCTs (99) 2 RCTs (99) IPT versus antidepressants IPT versus placebo IPT antidepressants versus antidepressants IPT IPT versus antidepressants IPT placebo versus medication clinic placebo 1 RCT (52) (1) Frank1990 254 (1) 98/77 (1) 40 (1) 10 week remission from MDD (100%). 14% bipolar (1) Placebo (1)–(2) 3 years (1)–(2) Not reported (1) 3 years (1) Not reported IPT versus IPT antidepressants No.

clinical management (1) 16 weeks (1) 5 months (1) 3 years (2) 2 years (1) Not reported (1) 6 months (1)–(2) Not reported Length of treatment High-intensity psychological interventions 255 (1) 16 weeks (2) 2 years (3) 6 weeks Follow-up (1)–(3) Not reported (1) Not reported . imipramine (98 mg/day) (1) 16 weeks (1) Nortriptyline. placebo (2) Paroxetine (10–40 mg/day). placebo.2 mg/day). trials (total participants) 3 RCTs (141) Study IDs (1) Reynolds1999 (2) REYNOLDS2006 (3) Weissman1992 (1) 80/75 (1) 80/75 (1) 99/69 (1) VAN SCHAIK2006 (1) Reynolds1999 (2) REYNOLDS2006 (1) 80/75 (2) 129/66 (1) 68 (2) 77 (1) 100% depressive disorder (1) GP care (1)–(2) 100% MDD N/% female (1) 80/75 (2) 129/66 (3) 25/71 (1) 68 (1) 68 (1) 68 Mean age (1) 68 (2) 77 (3) 70 (1) 100% MDD (1) 100% MDD Diagnosis (1)–(2) 100% MDD (3) 100% moderate/ severe MDD (1) Nortriptyline Comparator (1) Nortriptyline (1) Nortriptyline (2) Paroxetine (10–40 mg/day) (3) Alprazolam (2.Table 50: Summary study characteristics of IPT for the elderly IPT for the elderly IPT antidepressants versus antidepressants 1 RCT (46) (1) Reynolds1999 (1) Reynolds1999 1 RCT (45) 1 RCT (143) 2 RCTs (223) IPT versus antidepressants IPT versus standard care (Netherlands) IPT as maintenance treatment (2/3 years) IPT versus IPT antidepressants No.

0. The evidence indicated a significant effect in self-reported depression scores at post-treatment (SMD –0. 95% CI 0.73. Furthermore. 4.5. Two studies.60. In addition.73.28) and 9 months’ (SMD –0.16). Interpersonal therapy as a continuation treatment The evidence of one study (Schulberg1996) showed a small to medium significant effect (SMD –0. Similarly.07 and RR 0. 0.57. 95% CI 1. 0. 95% CI 0. Table 52. 0.73.81. Only one study.15) for IPT in reducing depression scores after 4 months’ continuation treatment when compared with treatment as usual.79. respectively. 95% CI 1. 95% CI. 0. 1. SCHRAMM2007.21) and 9 months’ (SMD –0. 0.32. 95% CI 0. 0.69. 95% CI 0. HRSD post-treatment SMD 0. and these results are reported in Section 8. Furthermore. However.43.56. and it is important to take this into consideration when interpreting the results.44.05) in the reduction of clinician-rated depression scores post-treatment. 8.13) followup.4 Clinical evidence summary Three studies included a comparison of IPT with CBT. there was a significant small to medium effect (SMD –0. 0.08. the study had still not been published when the guideline update was being prepared. The data for the Freeman2002 study is unpublished data the review team for the previous guideline obtained from the authors in anticipation of it being published. showed that when combination treatment was compared with antidepressants alone there was a significant medium effect (SMD 0. one study. 0.15. there was a large effect for IPT in reducing self-report depression scores at 3 months’ (SMD –0.3 Clinical evidence Evidence from the important outcomes and overall quality of evidence are presented in Table 51. Elkin1989 (n 123) looked at IPT when compared with placebo.98.High-intensity psychological interventions 8. 256 . 95% CI 1. examined the effectiveness of IPT versus antidepressants alone.22.5. 95% CI. 0.99). Elkin1989 and Schulberg1996.33.32. 95% CI 0.88.48. 95% CI 0.30). There was a higher risk for discontinuation in the placebo group when compared with IPT (RR 0.03. 0.40.04.75.97).1. Based on the evidence of one study (Reynolds1999) the combination treatment of IPT plus antidepressants when compared with IPT alone had a significant difference in decreasing clinician-rated depression scores (RR 2.37) follow-up. 0. Table 53 and Table 54. 0. 95% CI 1. in clinician-rated depression reports there was a large effect at 3 months’ (SMD –0. Four studies looked at IPT compared with usual GP care (including medication). The evidence showed no significant differences among the two groups (for depression scores: BDI post-treatment SMD 0. The full evidence profiles and associated forest plots can be found in Appendix 16b and Appendix 19b.93) for IPT in reducing clinician-rated depression scores at post-treatment when compared with placebo.26. 95% CI 1.40.41.

97) K 1.03 to 4.16 ( 0.46) 118 344 RR 1.02 SMD 0. 123 K Moderate Moderate Low Low Low K K 2. n 2.41 to 0.05 SMD 0. n IPT 16.22) IPT (with/without placebo) versus IPT antidepressants IPT antidepressants versus antidepressants IPT versus antidepressants IPT versus placebo Leaving study early for any reason RR 0.04 K 2.44 to 0.69 ( 1.15 to 0.01 & IPT 03.01 IPT 05.14) RR 0.05 SMD 0. n IPT 10. n 58 K 4. n Forest plot number IPT 02.94 (0.22 to 0.72 to 1.05 & IPT 17.Table 51: Summary evidence profile for IPT IPT versus usual GP care (including antidepressants) RR 3.64 to 0.12 (0.31 (1.30) Low 2.02 .94 to 5.79 to 0. n 302.12) HRSD at 12 weeks: SMD 0.07) Quality Low Number of studies.53 to 1.57 (0. n IPT 15.28) IPT 09.18) – HRSD 7: RR 2.01 33 Low Moderate K K 2.04 200.04 HRSD at 5/6 weeks: SMD 0.01 Depression self-report SMD 0. n 302 K Moderate Moderate Moderate 3.13 ( 0.63) Moderate 123 IPT 04. 87 130 Not reported SMD 0.04 ( 0. n 2. n IPT 14. n IPT 14.28 measures at endpoint ( 0.26 (1.03 IPT 11. participants K 1.30) HRSD 7: RR 1.05 High-intensity psychological interventions 257 Number of studies.07 ( 0.40) Low K 1.01 Depression clinicianreport measures at endpoint SMD 0. n 250 K 1. participants K 1.86) RR 0.33 to 0.56 to 0.93) Quality 123.08 ( 0. n 232 K 2.02 SMD 0.44 (0.43 ( 0.55 to 0. 160 IPT 16.16) Moderate 123 IPT 05. n Forest plot number IPT 01.06 ( 0.33 to 0. participants K K 1. n 72 – K – Low 1. n Forest plot number IPT 02.86 to 1.32 to 0.77 (0.07) HRSD 7: RR 0. n 1.73 (0.72 to 2.99) Quality Moderate Number of studies.

42 (0.06 IPT 20.06 Not reported – – – – Not reported IPT 19. 184 Forest plot number IPT 19.34) Low K 1.06 & IPT 18. n 1. n 16 – RR 5.22) Low K 1. participants K K 1.15) HRSD 7 after 4 months’ treatment: RR 0. 185 K 1. participants – Forest plot number – .06 – RR 0.04 (0.57 ( 1.73 to 0.53 to 0.02 to 9.66 (0.17 (0. n 23 – Low – – – – – – RR 0. n 1.27) Not reported Not reported Not reported IPT antidepressants versus antidepressants medication clinic IPT antidepressants versus IPT placebo IPT placebo versus medication clinic placebo 258 SMD 0.37) Quality Low Low Number of studies.06 IPT versus antidepressants High-intensity psychological interventions Depression clinician-report measures at endpoint SMD 0.06 Relapse Not reported Quality – Number of studies.44 ( 0.Table 52: Summary evidence profile for IPT as continuation treatment (up to 6 months) IPT as continuation treatment (up to 6 months) IPT versus TAU IPT antidepressants versus antidepressants After 6 months maintenance: SMD 0.79 to 1.50 (0.03 ( 0. n 184.82) Moderate Moderate K K IPT 19.26 to 115. n 185.06 IPT 20.41 to 0.26 to 0.01 to 3.32) HRSD 7 after 4 months’ treatment: RR 1. n 25 1.51) Low K 1.06 & IPT 18. n 15 – – IPT 20.

62 to 1.07 259 .07 RR 0.11 (0. 51 K n 2. 51 K n 2.80 (0. 102 Moderate K n 2.84 to 1.24 (0. 52 IPT 22. participants K n 1. 54 K n 1.22) RR 0. 49 K n 2.35 (0.06 to 1.18) Low K n 1.01) Low K n IPT 21.86 (0.00 to 2.07 IPT placebo IPT versus versus IPT medication placebo clinic placebo IPT antidepressants versus medication clinic antidepressants Not reported IPT versus IPT IPT antidep.antidepresressants sants versus IPT placebo Leaving study RR 0.02) Moderate K n 2.22 (0.Table 53: Summary evidence profile for IPT as maintenance treatment (3 years) IPT as maintenance treatment ( 3 years) IPT versus antidepressants IPT versus placebo IPT antidepressants versus antidepressants IPT antidepressants versus medication clinic placebo RR 2. 54 K n 1.89 (0.35 to 2.07 IPT 21.10 to 2.73 (1. 103 IPT 21. 101 Forest plot number IPT 21.07 Relapse RR 1.07 IPT 21.07 1.07 IPT 22. 106 IPT 22.48 early for any (0.10 to 0.27 to 0.74 to 7.26 to 1.50 (0.65) Quality Moderate High Number of studies.07 1.62 (0.01) Not reported IPT 21.87) Moderate K n IPT 21.42 (0.99) Low K n IPT 22.38 to 1. 106 Low Moderate RR 0.07 IPT 22.50) Low K n 1.57 to 1.28) Quality Low Moderate – – Number of studies.07 RR 1.65 to 6.40) reason RR 0.44) RR 0.66 to 0.07 – RR 0.98) RR 0.07 2. 103 IPT 22. participants K n 1.60 (0.07 RR 0.59 (0.11 to 3.49) Moderate – K n – 1.97) High K n 2.38) RR 0. 54 IPT 22.07 RR 2. 49 Low – RR 0.76 (0.10 to 2.07 RR 0. 101 High-intensity psychological interventions Forest plot number IPT 22.29 (0. 52 IPT 21.

participants K Forest plot number IPT 23.67) Low 121 IPT 24.03 to 4.87 (0.97) Quality Moderate Number of studies.52 to 1.44 to 0. n IPT 25.08 143.08 HRSD 7: RR 0.45) Quality Moderate Number of studies.63 (0.11 ( 0.61 to 0.66) Low 33 IPT 24.19 to 2.08 Depression clinicianreport measures at follow-up Not reported Quality – Number of studies. n 42 – Low – RR 0. participants – Forest plot number – .01 to 1.26 (1.Table 54: Summary evidence profile for IPT for the elderly IPT for the elderly IPT antidepressants versus antidepressants RR 0.05) MADRS at 6 months: SMD 0. n 41 K 1.08 – – IPT versus IPT antidepressants Leaving study early for any reason RR 0. n IPT 26.08 High-intensity psychological interventions Depression clinician-report HRSD 7 measures at endpoint RR 2.08 – Not reported – – – Not reported MADRS at 2 months: SMD 0. participants K Forest plot number IPT 23.30 to 1. n 41 K 1.94 to 2.10 (0.08 Not reported K 1.75) IPT 25. n Low 42 HRSD 7: RR 1.10) Not reported IPT versus antidepressants IPT versus standard care (Netherlands) 260 3. n 1. n 1.60 (0.71 (0.28 ( 0.22) – – – – Low K K 1. 143 K 1.

34) and also when compared with medication clinics (RR 5.26. Relate) tends to draw on these models. Interpersonal therapy as maintenance treatment Only two studies included a comparison of IPT as a maintenance treatment (Frank1990 and Reynolds1999B). 95% CI 0. 95% CI 1.02. Details on the methods used for the systematic search of the economic literature are described in Chapter 3. However. antidepressants had a significant effect in reducing clinician-rated depression measures (RR 1.17. Section 3. This significant effect was also seen when combination treatment was compared with antidepressants (RR 0. 95% CI 0. These conditions include the provision of positive regard. 1996). 95% CI 0. there are differences in how they are used 261 .49) when compared with medication clinics. 95% CI 0.22).High-intensity psychological interventions Based on the evidence of two studies with a continuation time of 3 years (Frank1990. Rogers’s original model was developed into structured counselling approaches by Truax and Carkhuff (1967) and. personalising and action. 2. favouring combination treatment of IPT plus antidepressants when compared with IPT alone. independently.50. 3.6 8. 95% CI 0.22.6.42.65) when compared with IPT plus placebo and (RR 0. n 42). n 33) indicated a significant effect (RR 2. Interpersonal therapy for the elderly The evidence for IPT in an elderly population is based on four studies (n 284). 8.1. Reynolds1990) the evidence indicates that combining interpersonal therapy and antidepressants has a lower risk of relapse when compared with IPT plus placebo (RR 0.01. genuineness and empathy. Voluntary sector counselling training (for example.6. 8. When IPT was studied as a maintenance treatment.94. 0.26.5. 9. although many other therapies now use the basic ingredients of clientcentred counselling (Roth & Fonagy.10. Based on the same study (Reynolds1999.42.27.5 Health economic evidence and considerations No evidence on the cost effectiveness of IPT for people with depression was identified by the systematic search of the economic literature. by Egan (1990) who developed the three stage model: exploration.60.97) for reducing clinician-rated depression scores in an elderly population. One study (Reynolds1999. combination treatment had a significant effect in lowering the risk of relapse (RR 0. 4. 95% CI 0.75) when compared with IPT. 0.1 COUNSELLING Introduction Counselling was developed by Carl Rogers (1957) who believed that people had the means for self-healing. problem resolution and growth if the right conditions could be created. 115.51).03.

References for studies from the previous guideline are in Appendix 18. GOLDMAN2006 and GREENBERG1998. Ward2000 was excluded because it did not meet inclusion criteria: only 62% met diagnosis for depression and this study was not completely randomised. 8. Rogers. are not listed in Table 55 because these compare two different types of counselling. GREENBERG1998. 8. respectively. a nonRCT) was examined.4 Clinical evidence summary One study (Bedi2000) compared the effectiveness of counselling versus antidepressants.. but are described in the text below.6. with a greater sense of well-being’. The content of these various approaches may include psychodynamic. 262 . The full evidence profiles and associated forest plots can be found in Appendix 16b and Appendix 19b. as this study was included in the previous guideline a separate sub-analysis has been conducted to determine whether this would have affected the GDG’s conclusions. discover and clarify ways of living more resourcefully. The results of this sub-analysis do not appear in the tables. 1986) and counselling has become a generic term used to describe a broad range of interventions delivered by counsellors usually working in primary care. Ward2000) were reported in the previous guideline. 2003). Three studies (Bedi2000.6. WATSON2003) meeting the inclusion criteria were found in the update search. Summary study characteristics of the included studies are in Table 55.2 Studies considered65 Three new studies (GOLDMAN2006. although some differences in the baseline scores of the patient preference group 65Study IDs in title case refer to studies included in the previous guideline and study IDs in capital letters refer to studies found and included in this guideline update. Definition The British Association for Counselling and Psychotherapy (BACP) defines counselling as ‘a systematic process which gives individuals an opportunity to explore. with full details in Appendix 17b.3 Clinical evidence Evidence from the important outcomes and overall quality of evidence are presented in Table 56. 2006. However. two of which are included in the guideline update.High-intensity psychological interventions (Kahn. which also includes details of excluded studies.. 8. Simpson2003. systemic or cognitive behavioural elements (Bower et al. A further trial (Stiles et al.6. but it was ultimately excluded because not all patients met criteria for depression and there were concerns about the selection of the study population. Two studies. 1985.

0. trials (total participants) Study IDs N/% female Mean age Diagnosis Comparator 1 RCT (103) (1) Bedi2000 (1) 79/77 (1) 39 (1) 100% MDD (1) Antidepressants (choice of three antidepressants and continued for 4–6 months) (1) 8 weeks (1) Not reported Counselling GP Counselling care versus GP care versus CBT 1 RCT (145) (1) Simpson2003 (1) 116/82 (1) 43 1 RCT (62) (1) WATSON2003 (1) 66/67 (1) 41 (1) Depressed criteria (1) 100% MDD (14–40 on BDI) (1) GP care (1) CBT Length of treatment Follow-up (1) 6–12 sessions (1) 12 months (1) 16 weeks (1) None suggest caution in interpreting the data. One study (Simpson2003) compared the combination of counselling plus GP care with usual GP care.15) but no significant differences between the two treatment groups on discontinuation and self-report depression scores at follow-up.33. A sub-analysis was conducted on Ward2000. There were no significant differences and the evidence remains inconclusive (self-reported depression scores at endpoint: SMD 0. 95% CI 0. clinician-rated depression scores at endpoint: RR 1.08.38.41.40). The results of this study should be treated with some caution as the introduction of a patient preference element to the trial led to considerable differences in baseline severity measures between the two arms. 95% CI 0. clinician-reported depression scores were significantly reduced in the antidepressant group when compared with counselling (RR 1.47 and at 12-month follow-up: SMD: 0. The results indicated a significant medium effect in self-report depression scores at posttreatment (SMD 0.83. The comparison of counselling versus CBT was included in one study (WATSON2003). This caution is support by the 12-month follow-up data. 1. 0. 95% CI 1. 95% CI 0.06.66.80.32.03.29. 0. 0.17.49. 95% CI 0. This study included a comparison of counselling versus GP care. 0. There was no evidence of any important clinical benefit of counselling plus GP care (BDI at 6 months: SMD 0.40 and at 12 months: SMD 0.20. There is insufficient evidence (only one small-sized study with 263 . 95% CI 0.81) and does not support a conclusion that counselling and antidepressants are equivalent. which did not meet the inclusion criteria but was raised during the consultation process.04. 1.83). 95% CI 0.High-intensity psychological interventions Table 55: Summary study characteristics of counselling Counselling versus antidepressants No.

03 ( 0. participants Forest plot number Depression selfreport measures at follow-up – – – At 12 months: SMD 0. n 145 Counselling versus CBT Not reported – – – SMD 0.94 (0.39 to 0.04 ( 0.43 to 2.02 C 17.73 to 1.17 ( 0.40) BDI 14: RR 0.07 ( 0. participants Forest plot number Depression selfreport measures at endpoint Quality Number of studies.32 to 0.02) Low Low Low Low – – – Not reported Quality Low – 264 .33 to 0. participants Forest plot number – – – SMD 0.40) BDI 14: RR 0.62 to 1.33 to 0.22) At 12 months: SMD 0.47) Low K 1.20) Low K 2.13 (0.29 to 0.66) – – – At 6 months: SMD 0. n 83 Counselling GP care versus GP care RR 1.06 ( 0.05 Not reported – – – Not reported C 02.High-intensity psychological interventions Table 56: Summary evidence profile for counselling Counselling versus antidepressants Leaving study early Not reported for any reason Quality Number of studies.80 (0. n 215 C 08.95) Low K 1.08 Not reported Depression clinician.Not reported report measures at endpoint Quality Number of studies.

0.31.05. 0.14. n 130. n 1.) Two studies (GOLDMAN2006 and GREENBERG1998). participants Forest plot number Low K 1.83) Quality Number of studies.High-intensity psychological interventions Table 56: (Continued) Counselling versus antidepressants Number of studies.21 and at 12-month follow-up BDI: SMD 0.05 Not reported – Not reported Depression clinician.04. compared two different types of counselling (and therefore are not included in the tables above). Some practice-based evidence was also reviewed (Stiles et al. 0. 1.02 C 10. 145 Counselling versus CBT – Forest plot number C 04.05. 95% CI 0..47). A smaller practice-based study (Marriott & Kellett.57.02 wide CIs) to reach any definite conclusion about the relative effectiveness of these two treatments (for BDI scores post-treatment: SMD 0. n 1.38.64. 0. other diagnoses were included in this study. 95% CI 0. 95% CI 0. These two studies are small in size and therefore results should be interpreted with caution. 2006) but the number of patients with depression in the study fell below the cut-off for inclusion. 95% CI. n 65 Counselling GP care versus GP care K K K K 1. 2009).RDC 3 at report measures 12 months: RR 1. C 09. Furthermore. The results favoured emotioned-focused therapy (BDI scores: SMD 0. 115. n 103 – – – – – – C 05. The participants in the trials included in this review were predominantly drawn from groups in the mild-to-moderate range of depression (mean baseline BDI scores between 18 and 26) and two trials included people with minor depression (BDI scores starting from 14) (Bedi2000 and Ward2000). (Individual outcomes for Ward2000 are: at endpoint BDI: SMD 0.02. Overall the evidence for counselling is very limited.08 to 1. C 09.33. C 10. participants K 1. GOLDMAN2006 compared client-centered counselling with emotion-focused counselling.07.20). 145. 265 .29).05. 95% CI 0.48. The evidence indicates that there was no significant difference between treatments in reduction of self-reported depression scores (SMD 0. This was still the case when a sub-analysis including the Ward2000 study was conducted (SMD 0. 0.38. 95% CI 0.82).13. n 1.04. 0. GREENBERG1998 examined the effectiveness of client-centered counselling versus process-experiential counselling.41 at follow-up (1.

. Sixty five patients were randomised to either treatment modality while a further 183 patients who chose their treatment modality were also analysed. overall. outpatient consultations and medications) and non-health service costs (lost productivity. 67This is the economic analysis of Bedi2000. cognitive analytic therapy and CBT but it was small and underpowered and it was not possible to reach any conclusion on the differential effectiveness of the treatments. 2000. However. GREENBERG1998 and GOLDMAN2006 adopt a process-experiential/emotionfocused model. 8.High-intensity psychological interventions which included only 34% with a diagnosis of depression. 266 . Over 9 months. for counselling and GP care. GP consultations and medications) were estimated. compared counselling. which is compared in the latter two trials with the Rogerian clientcentred model. The authors concluded that counselling in primary care was not cost effective in the short-term if indirect costs were taken into account but that. the study was effectively a costminimisation analysis. respectively. referral to counselling was no more clinically effective or costly than GP care.6.. Section 3. Simpson et al. The primary outcome of the clinical analysis was change in BDI scores.6. Friedli and colleagues (2000)66 compared non-directive counselling with usual GP care in a UK RCT of 136 people with referral symptoms being caused by depression and anxiety disorders (50% were given a GP diagnosis of depression). as no differences in clinical outcomes were detected between the two groups. Miller et al. The primary outcome measure used in the analysis was change in BDI scores. For example. counselling. outpatient. no significant differences were 66Note that this study was excluded from the analysis of clinical effectiveness as only 50% might have met diagnostic criteria for depression.1. However. In addition to the limited data available for counselling. Evidence tables for all health economics studies are presented in Appendix 15. whereas the studies by WATSON2003.5 Health economic evidence and considerations Three studies were identified in the systematic literature review that evaluated the cost effectiveness of counselling for people with depression and other common mental health problems (Friedli et al. travel and childcare) were estimated over this period. Bedi2000 and Ward2000 follow a Rogerian client-centred model of counselling. Simpson2003 a psychodynamic model. Miller and colleagues (2003)67 compared counselling with antidepressants in patients with major depression who were recruited from general practice. The time horizon of the analysis was 9 months and direct NHS costs (hospital inpatient stay. The time horizon of the analysis was 12 months and direct NHS costs (inpatient. 2003. Details on the methods used for the systematic search of the health economics literature are described in Chapter 3. interpretation of the results is complicated by the different therapeutic models adopted in the studies.. 2003). total direct NHS costs were £309 and £474 per person while total non-health service costs were £809 and £469 per person.

the study was effectively a cost-minimisation analysis. no significant differences in total mean costs per person were detected between the two randomised groups while the non-randomised counselling group was significantly more costly than the non-randomised antidepressant treatment group over 12 months. transference and counter-transference). However. The authors suggested that counselling might be a more cost-effective intervention in patients with mild to moderate depression but. Clinical trials of psychodynamic psychotherapy have focused on short-term psychological therapy (typically 10 to 30 weeks) usually in comparison with antidepressants or CBT. It is this brief version of psychodynamic psychotherapy. originating in the past.1 SHORT-TERM PSYCHODYNAMIC PSYCHOTHERAPY Introduction As with other schools of psychological therapy there are a number of variations on the original model of psychodynamic psychotherapy with some approaches focusing on the dynamic of drives (for example. Overall. This leads to patients being given an opportunity to explore feelings and conscious and unconscious conflicts. since there were no significant clinical differences detected between the two treatment groups. Simpson and colleagues (2003) evaluated the cost effectiveness of short-term psychodynamic counselling compared with routine GP care in a UK RCT of 181 patients with a history of depression.7 8. 267 . The primary outcome measure used in the clinical analysis was change in BDI scores. 2001). aggression) while others focus on relationships (Greenberg & Mitchell. Overall. for the larger patient group. with a technical focus on interpreting and working through conflicts. and where: ● Therapist and patient explore and gain insight into conflicts and how these are represented in current situations and relationships including the therapeutic relationship (for example. hospital. there was no statistically significant difference between the two treatment groups in total costs per person over 12 months (£1046 versus £1074). Other forms of this therapy have been influenced by attachment theory (Holmes.High-intensity psychological interventions detected between the two treatment groups at 12 months. 1983). which is the focus of this review. The authors suggested that there was no cost-effectiveness advantage of counselling over routine care for general practice attendees with chronic depression. The time horizon of the analysis was 12 months and direct healthcare costs (specialist mental health. primary care and community health and social care services) were estimated for this period. Definition Psychodynamic psychotherapy is defined as a psychological intervention derived from a psychodynamic/psychoanalytic model. often referred to as short-term psychodynamic psychotherapy. 8.7. antidepressant treatment was likely to be the more cost-effective intervention.

3 Clinical evidence Evidence from the important outcomes and overall quality of evidence are presented in Table 58.4 Clinical evidence summary Problems with unextractable data and multiple different comparators limited the analyses it was possible to undertake for this review. Summary study characteristics of the included studies are in Table 57. which also includes details of excluded studies.03. trials including participants without a diagnosis of depression and authors replacing dropouts. therefore. One study (Guthrie1999) was not included because in the sample population (which was selected on the basis of high attendance at outpatient clinics) only 73.1. compared short-term psychodynamic psychotherapy with behavioural activation and this is reported in Section 8. References for studies from the previous guideline are in Appendix 18.7.2 Studies considered68 In total.7. Two studies (Gallagher-Th1994 and Shapiro1994) included a comparison of short-term psychodynamic psychotherapy with CBT.7. However. papers not reporting outcome data. Reasons for exclusion included: trials not being RCTs. 95% CI 0.6% met diagnosis for depression and.2. 8. 8. did not meet the inclusion criteria (which is 80% of the total population). while it is not included in the main analyses and tables. The full evidence profiles and associated forest plots can be found in Appendix 16b and Appendix 19b.High-intensity psychological interventions ● Therapy is non-directive and recipients are not taught specific skills (for example. One study (McLean1979). re-evaluating. a sub-analysis including this paper was conducted and is reported below. the results of which are reported in Section 8.43. typically 16 to 20 sessions but with a range of 10 to 30 sessions across the included studies. The evidence from one study (DEKKER2008) showed a significant medium effect (SMD 0. or problem-solving). 8. 268 . 0. respectively. with full details in Appendix 17b. 17 studies were found in the search for trials of short-term psychodynamic psychotherapy. thought monitoring. It should be noted that all the included studies were of short-term psychodynamic psychotherapy and therefore the analysis and subsequent recommendations are limited to short-term psychodynamic psychotherapy.82) favouring antidepressants when compared with short-term psychodynamic 68Study IDs in title case refer to studies included in the previous guideline and study IDs in capital letters refer to studies found and included in this guideline update. Ten studies were included (six were found in the update search and four were also reported in the previous guideline) and seven were excluded.

Table 57: Summary study characteristics of short-term psychodynamic psychotherapy STPP antidepressants versus supportive therapy antidepressants 1 RCT (74) (1) Burnand2002 (1) DEJONGHE 2004 (1) MAINA 2005 (1) MAINA 2005 1 RCT (191) 1 RCT (20) 1 RCT (20) 2 RCTs (220) (1) KOOL2003 (2) MAINA2008 STPP versus STPP antidepressants STPP versus waitlist control STPP versus supportive therapy Antidepressants versus STPP antidepressants Short-term psychodynamic psychotherapy (STPP) versus antidepressants No.up (1) Not reported (2) 3 months (3) 4 months (1) Not reported (1) 6 months (1) 6 months (1) Not reported (2) 48 months . trials (total participants) 3 RCTs (230) Study IDs (1) DEKKER2008 (2) McLean1979 (3) SALMINEN 2008 (1) 45/61 (1) 128/67 (1) 19/ 63 (1) 19/63 N/% female (1) 76/74 (2) 111/72 (3) 35/68 (1) 36 (1) Unextractable (1) 37 (1) 37 (1) 79/62 (2) 56/61 (1) 34 (2) 36 Mean age (1) Unextractable (2) 39 (3) 42 (1) 100% MDD (1) 100% MDD Diagnosis (1) 100% depressive episode (2) 100% MDD (3) Mild/ moderate episode of MDD (1) Clomipramine (125 mg/day) (1) GP’s choice of antidepressant (1) 100% dysthymia or subthreshold depressive symptoms (1) 100% dysthymia (1) 100% or subthreshold depressive episode depressive symptoms (2) Remission from MDD (follow-up of MAINA2005) (1) Waitlist control (1) Supportive psychotherapy (1) Range of antidepressants (2) Citalopram/ paroxetine (20–60 mg/day) (1) 15–30 weeks (1) 15–30 weeks (1) 24 weeks (2) 6 months Comparator (1) Venlafaxine (225 mg/day) (2) Amitriptyline (150 mg/day) (3) Fluoxetine (1) 10 weeks Length of treatment High-intensity psychological interventions (1) 24 weeks (2) 10 weeks (3) 16 weeks (1) Not reported (1) Up to 6 months 269 Follow.

04 ( 0.43 (0. n PP 14.03 1.07 20 Very low K 1. n 57 – – – – – SMD 1.01 Depression selfreport measures at endpoint Not reported Quality – Number of studies.76) Low K 1.( 4.97 ( 1.97) Low K PP 09.51 to 1.03 Still meeting WMD 0. n 208 – Moderate – RR 0. n 103 Forest plot number PP 02.44 to 2.01 .82) Quality Moderate Number of studies.43 (0.97 to 2.01 to 0.05 – Not reported 66 K 1.90 (0. n 193 High-intensity psychological interventions Forest plot number PP 01. n 95 K 1.32) 1.03 SMD 0.06 20 Moderate K 1.91 to 0.46) treatment: RR 1.35 ( 0.09 ( 2.04 Not reported Not reported Not reported PP 15.71 to 2.80 RDC criteria post.70 (0.16 ( 2.13) Not reported SMD 0. n PP 18.89) Low 1.08 128 PP 13.23 to 0. n PP 11.60) Quality Moderate Number of studies.16 (0. participants – Forest plot number – Depression clinician-report measures at endpoint SMD 0.05 208 Moderate K 1.04 STPP versus antidepressants Leaving study early for any reason RR 0.30) Low K PP 06.06 to 2. 66 K 1. n Low 74 – SMD 0.04 to 0.03) SMD 0. participants K 1. n PP 16. n PP 17.61 to 1.76) Low K PP 05. participants K 2.81 to 5.06 (0.44) Not reported Not reported Not reported STPP versus STPP antidepressants STPP versus waitlist STPP versus supportive therapy STPP antidepressants versus antidepressants 270 RR 2.03 to 0.Table 58: Summary evidence profile for short-term psychodynamic psychotherapy STPP versus CBT STPP antidepressants versus supportive therapy antidepressants RR 1.

95% CI 0. 0.10.91.62. 2.50). 95% CI 0. 95% CI 0. 95% CI 1.52.62.5 Health economic evidence and considerations One study (Guthrie et al. 1. 1999) was identified in the systematic literature review that evaluated the cost effectiveness of short-term psychodynamic psychotherapy for people who are high utilisers of psychiatric services (with 73.52. 95% CI 2. 95% CI 0.81.15. 95% CI 1. Some of the difficulty derives from there being a number of different comparators in a small dataset.59. However.58). one study (MAINA2005) showed a significant and large effect (SMD 1. 0. KOOL2003 compared short-term psychodynamic psychotherapy in two different populations with depression: one with comorbid personality disorder and the second without.High-intensity psychological interventions psychotherapy in the reduction of clinician-rated scores at endpoint.97. 0. 95% CI 1. this is a weak dataset characterised by a number of the findings being contradictory and/or difficult to interpret.09. but given the wide CIs and size of the study it is difficult to establish a clear picture of this comparison. 0.6% having a diagnosis of depression). 95% CI 0. There is limited evidence for a benefit of short-term psychodynamic psychotherapy (typically 16 to 20 sessions over 4 to 6 months) in a population with dysthymia and subthreshold depressive symptoms over waitlist or usual care and inconsistent findings when compared with antidepressants.16. When compared with a waitlist control.69 and SMD 1.7. 0.07.95) and at 48 months (SMD 0. favouring short-term psychodynamic psychotherapy. 95% CI 0.50.24.04. A follow-up study (MAINA2008) showed that adding short-term psychodynamic psychotherapy to antidepressant treatment had a significant medium to large effect at 24 months (SMD 0. MAINA2005 and MAINA2008 were conducted in a population diagnosed with minor depression or dysthymia. 0. When a separate analysis was conducted with the Guthrie1999 study.22 and at 6-month follow-up SMD 0.53.01) in reducing clinician-rated depression scores when compared with antidepressants alone. Details on the methods used for the systematic search of the health 271 . 8. One study (McLean1979) indicated a significantly higher risk of discontinuation in those treated with short-term psychodynamic psychotherapy when compared with behaviour therapy (RR 3.03) for short-term psychodynamic psychotherapy in clinician-rated depression scores at post-treatment when compared with supportive therapy.03.18 respectively) but the small sample size in the population without personality disorder suggests caution when interpreting this result. The results suggest that short-term psychodynamic psychotherapy is more effective in people diagnosed with depression and personality disorder than those without (SMD 1. Comparisons against other active psychological interventions are also very limited. 8.13) in clinician-rated depression scores at post-treatment. This study also indicated a large effect (SMD 0. the evidence was inconclusive given the small size of the study and the wide CIs (for SCL-90-R at endpoint: SMD 0. In summary.13).16. the results of a further small-sized study (SALMINEN2008) showed no significant differences between short-term psychodynamic psychotherapy and antidepressants when looking at the mean change from baseline to endpoint (SMD 0.02. 0.

1 RATIONAL EMOTIVE BEHAVIOURAL THERAPY Introduction Rational emotive behavioural therapy is a form of CBT developed by Albert Ellis in the 1950s and 1960s (Ellis.21). outpatient. total societal costs per person were lower in the brief psychodynamic interpersonal therapy group at 6 months (US $1959 versus $2. There is an emphasis on addressing thinking that underpins emotional and behavioural problems. resulting in lower costs but better outcomes. The brief psychodynamic interpersonal therapy group also gained more QALMs during this period (4. 272 . p 0. 69Each study considered for review is referred to by a study ID in capital letters (primary author and date of study publication. comparison with CBT can be found in Section 8.8. The primary outcome measures used in the economic analysis were quality-adjusted life months (QALMs).1. and only one study met the criteria of the GDG (DAVID2008). While brief psychodynamic interpersonal therapy appeared to dominate treatment as usual. 8. Patients learn how to examine and challenge their unhelpful thinking.8 8.1.13).48.5% diagnosed with depression). relatively short-term therapy usually delivered one-to-one that uncovers and addresses the relationships between thoughts. day cases. which were estimated from utility weights derived from the EQ-5D questionnaire. The time horizon of the analysis was 6 months posttreatment and direct NHS costs (inpatient. feelings and behaviours.8. 8. 1962). Compared with CBT it has been subject to fewer research trials. Overall. This study compared rational emotive behavioural therapy with antidepressant medication. p 0.6. A&E visits and medications) and non-health service costs (travel and lost productivity) were estimated during this period. Definition Rational emotive behavioural therapy is a present-focused.2 Studies considered69 Only one RCT (DAVID2008) was found and was included in the review. neither the cost nor QALM differences between the two treatment groups were statistically significant. This section reports on the comparison of rational emotive behavioural therapy with antidepressants. then a date is not used). Section 3.87 versus 3. Evidence tables for all health economics studies are presented in Appendix 15. The study by Guthrie and colleagues (1999) compared brief psychodynamic interpersonal therapy (equivalent to short-term psychodynamic psychotherapy) with treatment as usual in a UK RCT of 144 patients with non-psychotic disorders (75.High-intensity psychological interventions economics literature are described in Chapter 3. except where a study is in press or only submitted for publication.465.

CI 0.37).63. a summary of the evidence profile is not included here.37 to 0. respectively. 8.07.8.5 Health economic evidence and considerations No evidence on the cost effectiveness of rational emotive behavioural therapy for people with depression was identified by the systematic search of the economic 273 . which also includes details of excluded studies.8.29. However. the findings were promising in terms of end-of-treatment depressive symptoms and in terms of acceptability (RR 0.4 Clinical evidence summary The evidence of one study (DAVID2008) showed no clinically important different effects of rational emotive behaviour therapy in depressed patients when compared with antidepressants (BDI: SMD 0.22 to 1.44 to 0. with full details in Appendix 17b.61).80) and preventing relapse at 6 months’ follow-up (RR 0. 8.00.8. HRSD: SMD 0.3 Clinical evidence Because of the small dataset. 95% CI 0.02 to 1.20. trials (total participants) Study ID N/% female Mean age Diagnosis Comparator Length of treatment Follow-up 1 RCT (180) DAVID2008 113/66 37 100% MDD Antidepressants 14 weeks 6 months Summary study characteristics of the included studies are in Table 59. 95% CI 0. CI 0.High-intensity psychological interventions Table 59: Summary study characteristics of rational emotive behaviour therapy Rational emotive behavioural therapy versus antidepressants No. 8. The full evidence profiles and associated forest plots can be found in Appendix 16b and Appendix 19b.

remission and relapse remained the same as reported in the previous guideline model. levels of depression severity in relation to the HRSD and BDI were based on those proposed by the APA (2000a) rather than those proposed in this guideline.2 Methods A pragmatic decision analytic model was constructed using Microsoft Excel XP. A detailed structure of the decision tree is presented in Figure 6. The following strategies were considered: Strategy A: Antidepressant treatment given for 12 weeks with 6 months’ maintenance therapy and 6 months’ follow-up (AD). However.9 8.High-intensity psychological interventions literature.9. However. it was necessary to update the economic model in order to better reflect current medical practice within the UK. Based on GDG expert opinion. 8. Strategy B: Combination of 12 weeks’ antidepressant treatment and 16 sessions of CBT with 6 months’ maintenance therapy and 6 months’ follow-up (COMB).9.1 ECONOMIC MODELLING Background The aim of this economic analysis is to update the model constructed in the previous guideline (NICE. This included 3 months of the initial therapy. A time horizon of 15 months was chosen to reflect the available comparative clinical evidence. Within the model patients either continue or discontinue their initial treatment. This included the additional costs of maintenance therapy in both treatment groups while other input parameters. evidence of the clinical effectiveness of IPT or behavioural activation compared with antidepressant treatment for moderate or severe depression was limited. Details on the methods used for the systematic search of the economic literature are described in Chapter 3. Patients in remission can then either relapse or remain in remission health states. in this economic analysis.6. Therefore.1. Clinical outcome data within the model including rates of discontinuation. Section 3. were also updated. followed by 6 months’ maintenance therapy and 6 months’ follow-up. 8. after which they enter remission or no remission health states. CBT was again chosen as the form of psychological therapy for this analysis as the clinical evidence was superior and CBT remains more widely available in the UK compared with other high-intensity interventions. It should be noted that these data were taken from meta-analyses that were undertaken in the previous guideline. It was anticipated that other high-intensity psychological interventions such as IPT or behavioural activation would be evaluated in an economic model. 274 . including patient utility scores and unit costs. 2004a). which evaluated the cost effectiveness of antidepressant treatment versus a combination of antidepressant treatment and CBT for the routine treatment of moderate/severe depression.

Unit costs were obtained from a variety of sources including the British National Formulary (BNF 56. 2008) and the Personal Social Services Research Unit (PSSRU) 275 . three specific strategies for the first-line management of depression were considered.High-intensity psychological interventions Figure 6: Structure of the model Relapse Remission Continue treatment No Relapse No Remission Patients with moderate/severe depression Relapse Remission No Relapse Discontinue treatment No Remission Originally. 8. the updated clinical evidence review showed no overall superiority for CBT alone on treatment outcomes over antidepressant treatment. The efficacy evidence combined with the significantly higher treatment cost of CBT compared with the cost of antidepressant treatment resulted in the exclusion of CBT alone from the final analysis. only direct health and social care costs were considered in the analysis. resource utilisation data were collected as part of the literature review or from GDG expert opinion. similar to the previous guideline. However.3 Model assumptions Population Two separate models were constructed for a hypothetical cohort of 100 patients in each treatment group with either moderate or severe depression. In order to cost the two therapy pathways. Therefore.9. 2008b). Resource use and unit costs An NHS and PSS perspective was taken for the analysis based on current NICE guidance (NICE.

Citalopram was used to represent standard pharmacotherapy for patients with moderate or severe depression because it was the most commonly prescribed antidepressant in 2007 in England (Department of Health. 2008). It was also assumed that patient monitoring in both primary and secondary care consisted of two fortnightly visits in the first month followed by one visit per month.. it was assumed that all patients with moderate depression and 50% of patients with severe depression would receive standard GP care while the remaining 50% of patients with severe depression would receive specialist mental health outpatient care (GDG expert opinion). However. Resource utilisation data were then combined with the relevant cost associated with each therapy. it was assumed that patients who discontinued initial treatment would drop out after 4 weeks of treatment. 2009). it was assumed that combination therapy would consist of 16 sessions of CBT over 12 weeks. it was assumed that no further additional treatment or mental healthcare resources beyond the 6-month maintenance period were required. One CBT session lasts for 55 minutes and is provided by a specialty doctor. As such.High-intensity psychological interventions (Curtis. it was assumed that both moderate and severely depressed patients would receive an additional two CBT sessions. irrespective of treatment group (Rush et al. while the maintenance therapy period consisted of one GP/specialist visit every 2 months (GDG expert opinion). Cost data for subsequent mental healthcare were taken from a study published by the King’s Fund which estimated annual mental healthcare costs for respondents with mild. in addition to the antidepressant treatment protocols described above (GDG expert opinion). 2009). it was assumed that they would continue to consume additional mental healthcare resources over the 15-month time horizon. 2006. As in the case of outcomes. 2008a). clinical psychologist or mental health nurse (Curtis. in addition to the antidepressant (AD) maintenance therapy protocols described above (GDG expert opinion). As part of patient monitoring. Combination therapy For both moderate and severely depressed patients. GDG expert opinion). Antidepressant treatment The antidepressant treatment protocol consisted of 12 weeks plus 6 months’ maintenance period of 40 mg of generic citalopram per day for both moderate and severely depressed patients (GDG expert opinion). no discounting was applied because the time horizon was 15 months. moderate and severe depressive disorder based on the UK psychiatric morbidity survey (McCrone et al.. 2009). For patients in remission who did not relapse during follow-up. these annual mental healthcare costs may be an under estimate of the actual costs incurred by patients with 276 . To revise costs downwards. for patients with unsuccessful treatment outcomes. During the 6-month maintenance therapy period. All costs were based on 2007/08 prices and were inflated where necessary using Hospital and Community Health Service indices (Curtis. Subsequent healthcare Patients who discontinued initial treatment did not incur the full costs of treatment.

only 65% of people with depression were in contact or receipt of mental health services. the baseline absolute rates of remission. The guideline meta-analysis of non-remission rates was based on intention-to-treat analysis. These annual costs were divided into monthly cost estimates and then projected for the periods during which unsuccessfully treated patients would consume subsequent mental healthcare estimated in the model. 30) would spontaneously enter remission (GDG expert opinion). According to the survey. The dichotomous outcome measure of no remission was defined by scores greater than six on the 17-item HRSD or more than eight on the 24-item HRSD. Therefore. Full details of event probabilities are presented in Table 62. as non-remitters). it was assumed that rather than remaining moderately or severely depressed.. For the sensitivity analyses. social services. these subsequent mental healthcare costs were weighted downwards based on the assumption that 35% of patients would not incur any further healthcare costs. with non-completers being considered as an ‘unfavourable’ outcome (that is. and was subsequently incorporated in the respective branch of the decision tree.High-intensity psychological interventions moderate and severe depression. discontinuation and relapse for antidepressant treatment as well as the respective relative risks of combination therapy versus antidepressant treatment were taken from the relevant guideline meta-analyses. For patients who did not complete their initial therapy. Full details of all resource use and unit cost parameters are presented in Table 60. Patients who did not achieve remission following therapy incurred full 3-month treatment costs followed by subsequent mental healthcare thereafter. No discounting of outcomes was necessary since the time horizon of the model was 15 months. the rate of relapse was estimated as 67% based on a study of patients who were not receiving maintenance antidepressant treatment (Murphy et al. For the economic analysis. 1984). this is likely to be an over estimate of the relapse rate for patients in this analysis who are receiving maintenance therapy. as one would expect respondents with mild depression to use less mental healthcare on average. 277 . For the base case analysis. it was assumed that the average time to relapse was based on the midpoint of the clinical relapse data elicited in the guideline meta-analysis. This means that non-remission rates included people who completed treatment but did not remit plus people who did not complete treatment. Therefore. a small proportion (20%: 95% CI 10. These two probabilities were applied to patients in both treatment arms. which was estimated over a 12-month period (GDG expert opinion). Clinical outcomes and event probabilities The outcome measure used for the economic evaluation was the quality-adjusted life years (QALYs) gained from either treatment. The key clinical parameter estimates – discontinuation rates. the proportion of non-remitters in the completer group was estimated from the available data. 95% CIs around the relevant relative risks of combination (COMB) therapy versus antidepressant (AD) treatment were used. remission rates and relapse rates – were collected as part of the updated clinical systematic review undertaken for the guideline. For patients in remission. GP visits and medication costs. residential care. For patients who relapsed while in remission. These mental healthcare costs included hospital and outpatient care.

16 sessions over 3 months plus 2 sessions during 6-month maintenance phase (moderate and severe) .Citalopram (40 mg/day): £18 3 months plus 6 months’ maintenance .Subsequent mental healthcare: 14 months AD total cost (moderate) AD total cost (severe) COMB total cost (moderate) COMB total cost (severe) £1044 CBT session (55 minutes): £58 (Curtis.Antidepressant treatment protocol (moderate) .High-intensity psychological interventions Table 60: Resource use and cost estimates applied in the economic model Resource use estimate Cost Source of unit costs Non-proprietary: £1.060 £2.Patient monitoring (severe): £581 50% mental health outpatient consultations ( 7) 50% GP consultations ( 7) over 9 months Total cost (moderate) £270 Total cost (severe) £599 Combination therapy (COMB) .Patient monitoring (moderate): £252 7 GP consultations over 9 months .. 2008) Continued 278 .AD: 1 month .2 GP consultations (moderate) .154 Monthly cost of subsequent mental health care: £165 Weighted by 65% according to proportion in contact with mental health services (McCrone et al. 2009) Antidepressant treatment (AD) .CBT sessions: 6 (moderate and severe) . 2009) Mental health outpatient consultation: £130 (Curtis. 2009) £270 £599 £1314 £1643 £2 £72 £166 £348 £1638 £1712 £1806 £2.50% GP care ( 2) 50% outpatient care ( 2) (severe) .87 per 28-tab pack (BNF.Antidepressant treatment protocol (severe) Total cost (moderate) Total cost (severe) Patients who discontinue treatment . 2008) GP consultation: £36 (Curtis.

4 GP consultations (moderate) .CBT sessions: 16 (moderate and severe) . 279 .Subsequent mental healthcare: 8 months AD total cost (moderate) AD total cost (severe) COMB total cost (moderate) COMB total cost (severe) Cost Source of unit costs £6 £144 £332 £928 £1404 £1554 £1742 £2. Utility scores represent the HRQoL associated with specific health states on a scale from 0 (death) to 1 (perfect health).Subsequent mental healthcare: 12 months AD total cost (moderate) AD total cost (severe) COMB total cost (moderate) COMB total cost (severe) Patients who relapse while in remission .CBT sessions: 17 (moderate and severe) .Antidepressant treatment: (3 6 months) .AD: 3 months .7 GP consultations (moderate) .50% GP care ( 4) 50% outpatient care ( 4) (severe) .482 £2.50% GP care ( 7) 50% outpatient care ( 7) (severe) . they are estimated using preference-based measures that capture people’s preferences and perceptions on HRQoL characterising the health states under consideration.345 Utility data and estimation of quality-adjusted life years In order to express outcomes in the form of QALYs.016 £2. the health states of the economic model needed to be linked to appropriate utility scores.High-intensity psychological interventions Table 60: (Continued) Resource use estimate Patients not achieving remission .670 £18 £252 £581 £1044 £702 £972 £1301 £2.

Peveler et al. 2000. McSad is a direct utility measure in which rating scale (RS) and standard gamble (SG) techniques were used to obtain utilities for specific health states. Pyne et al. 2005).. 2005). 1995). Bennett et al. In all three studies.High-intensity psychological interventions 8. ‘responder non-remitters’ and ‘non-responders’. patients with MADRS scores lower or equal to 12 were considered as ‘remitters’ and others considered as ‘non-remitters’.. 1998. 2000.. eight publications were identified that reported utility scores relating to specific health states and events associated with depression (Revicki & Wood.. Three studies used the EQ-5D Index instrument. Again. 1998. 2003). reported as median scores. Lenert et al. The other five studies used a variety of instruments to measure patient utility (Revicki & Wood. King et al. prospective cohort of patients with a new episode of major depressive disorder recruited in the French primary care setting assessed at 8 weeks’ follow-up. At 8 weeks. The study by Peveler and colleagues (2005) was another HTA based on an RCT comparing the cost-utility of TCAs. Peveler et al..4 Systematic review of published utility scores for adults with depression Among the studies already assessed for eligibility. utility scores improved from baseline at 12 months in all three treatment groups but no differences were detected between groups. EQ-5D utility scores were stratified according to depression severity (defined by CGI scores). 2004. no differences in median scores were detected between the three patient groups. SSRIs and lofepramine among UK patients with a new episode of depressive illness (based on GP diagnosis). The study by Sapin and colleagues (2004) was based on a multicentre. improved from baseline in all three treatment groups at 4 and 12 months. Bennett et al. However. These two patient groupings also led to the creation of three mutually exclusive groups: ‘responder remitters’. Schaffer et al. Sapin et al. preference values elicited from the UK population sample were used (Dolan & Williams. and by clinical response (defined by MADRS scores) at follow-up. currently recommended by NICE as a measure of patient HRQoL for use in cost-utility analyses (King et al.. Utility scores were 280 . Patients completed the EQ-5D questionnaire on a monthly basis over 12 months... to estimate utility scores for a cross-sectional sample of patients who had experienced at least one episode of major unipolar depression in the previous 2 years. 2000. 2000. 2002... Sapin et al. 2003.. King and colleagues (2000) collected patient EQ-5D utility data over 12 months’ follow-up in an RCT comparing usual GP care with two types of brief psychological therapy (nondirective counselling and CBT) among patients with depressive or mixed anxiety/depressive symptoms (BDI 14). Schaffer et al..9. Patients with a decrease of at least 50% in relation to baseline score were considered as ‘responders’ and others as ‘non-responders’. The health state classification system contains six dimensions (emotion/self-appraisal/cognition/physiology/behaviour/role-function). 2002.. Bennett and colleagues (2000) used a disease-specific measure. Lenert et al. 2004.. Pyne et al. Patient utility. the McSad instrument. 2000. each with four levels of dysfunction (none/mild/moderate/severe). 2000.

Cluster analysis was applied to the SF-12 HRQoL instrument to generate the six health states. Similarly. 8. were framed in terms of 1 month’s duration and described symptom severity. The resulting six-state health index model was then applied to HRQoL data taken from a longitudinal cohort study of patients with current major depression or dysthymia.High-intensity psychological interventions generated for three temporary clinical marker states of 6 months’ duration (mild/moderate/severe depression) and chronic states of lifetime duration (selfreported and severe depression). 1998. QWB-SA scores improved during follow-up for treatment responders (defined by a 50% reduction in HRSD-17 scores) but did not improve for non-responders. 2000). The individual symptom profiles each consisted of five statements describing a particular aspect of a symptom of depression. DSM–IV and Structured Clinical Interview for DSM-IV [SCID-IV]).. BDI. The scoring function of the QWB-SA was based on rating scale measurements taken from a random sample of the US population. The depression-related health states varied depression severity (mild/moderate/severe) and medication (nefazodone/fluoxetine/ imipramine). 281 . Pyne and colleagues (2003) used the self-administered Quality of Well-Being scale (QWB-SA) in a prospective cohort of US patients treated with antidepressants to measure change in patient HRQoL scores over 4 months’ follow-up.9. Bennett et al.5 Summary Table 61 summarises the methods used to derive health states and estimate utility scores associated with various levels of depression severity and treatments for depression as well as utility scores from each study. functioning and wellbeing. Utilities applied to the six health states were elicited through the use of visual analogue scale (VAS) and SG methods. MADRS. Overall. incorporating the content of several depression scales and interviews (HRSD. Revicki and Wood (1998) used SG techniques in US and Canadian patients with major depressive disorder to generate utility scores for 11 hypothetical depressionrelated and current health states according to depression severity and antidepressant treatment. as well as medication therapy including side effects. the study by Schaffer and colleagues (2002) used SG techniques to elicit utility scores for ten individual symptom profiles of major depression plus three ‘clinical marker’ depression profiles (mild/moderate/severe) among patients with current or past depression. the studies reviewed here reported significant impact of depression on the HRQoL of patients with depression. Lenert and colleagues (2000) estimated utility scores among depressed US primary care patients based on six health states according to level of depression severity (mild/severe) and physical impairment (mild/moderate/severe). A number of studies indicated that patients valued the state of severe depression as being close to zero (death) (Revicki & Wood. There was some limited evidence to suggest that generic utility measures such as the EQ-5D may be less sensitive than disease-specific measures such as the McSad health state classification system.

73 0.27) EQ-5D (TTO) 464 eligible patients with depressive symptoms 104 US depressed primary care patients VAS.29–0.74–0.73 0. SG Table 61: Summary of studies reporting utility scores relating to specific health states and events associated with depression Study Definition of health states High-intensity psychological interventions Bennett et al.85 Near-normal health (no depression) Mild mental with mild physical impairment Severe physical health impairment Severe mental health impairment (severe depression) Severe mental and moderate physical impairment Severe mental and physical impairment GP care 0.73 0.282 Valuation method 105 patients with history of major.07) ND counselling 0..34) 0. The health state descriptions referred to untreated depression.79 (0. King et al.85 0.81 0.66 (0.62) 0.13) 0. unipolar depression in the previous 2 years Temporary states (6-month): .85 0.83) 0.85 Population valuing Results (95% CI/SD) 0.81 (0.55–0. 2000 Cluster analysis used to obtain six health states from SF-12.21) 0.Moderate depression .94 (0.04 (0. 2000 RCT comparing three treatments: usual GP care and two types of brief psychological therapy (non-directive [ND] counselling and CBT) over 12 months’ follow-up..83 (0..78 (0. and those with no response.32 (0. The utility change scores over longitudinal study period were calculated using estimated health state utilities for those in remission.59 (0.18) 0. responder – non-remitters. 2000 Utility values were elicited using the SG McSad health state classification system. .Self-reported health state .22) 0.Mild depression .Severe depression Clinical states (lifetime): .87 (0. Lenert et al.09 (0.21) 0.85 0.20) 0.Severe depression CBT Baseline 4 months 12 months 0.05–0.01–0.

0.27) 0.21) -Fluoxetine 0.20) Depression remission.13) -Fluoxetine 0.64 (.70 (0.72 (0.55 (0. Depression response data (50% reduction in HRSD-17) collected at baseline.43 Pragmatic RCT of three classes of antidepressant: TCAs.54 0.7–21.58 (0. QWB-SA: 0. and the patient’s current health status.63 (0. 2005 EQ-5D (TTO) Baseline 12 months 261 UK primary care patients with new episode of depression 58 US patients treated for MDD Baseline (HRSD-17: 20.86 (0..23) -Fluoxetine 0.46 4 months 0. SG varying depression severity and antidepressant treatment.17) Remission.61 (0.20) -Imipramine 0.77 (0.22) Continued High-intensity psychological interventions 283 .78 (0.63 (0. maintenance treatment -Nefazodone 0.63 0.19) LOF 0.57 (0.19) SSRI 0. no treatment 0. SSRIs and lofepramine (LOF) over 12 months’ follow-up.21) Pyne et al. 70 patients with MDD from primary care practices in US and Canada Severe depression. 2003 Prospective observational study conducted over 16 weeks.28) 0.16) Revicki & Wood.19) -Imipramine 0. 4 weeks and 4 months. QWB-SA (Category scaling) TCA 0..03) Mild depression -Nefazodone 0. 0.78 (0.83 (0.73 (0. untreated Moderate depression -Nefazodone 0.43) Responders: Non-responders: 4 weeks 0. 1998 11 hypothetical depression-related states.80 (0.15) -Imipramine 0.41–0.Peveler et al.30 (0.27) 0. Treatment with antidepressant and/or mood stabiliser.

44 (0. prospective. poor concentration.27) 0.34) Controls 0.22) 0..79 (0.85 (0.21) 0. defined by MADRS scores.30 (0.13) 0.80 (0.86 (0. 2004 Multicentre.31) Past 0.15 (0.21) Baseline Mild depression Moderate depression Severe depression Population valuing Results (95% CI/SD) 250 patients with new episode of MDD not treated with antidepressant before inclusion. Clinical response. energy.36) 0.67 (0.47 (0. . 35 healthy controls) Mild Moderate Severe Current 0.69 (0.45 (0.72 (0. 2002 Utility scores for ten individual symptoms of depression and three depression severity profiles (mild/moderate/severe). 21 past depression. sleep.19) 0.20) 0.59 (0.24) 0. suicidal ideation): 0.28) SG 75 Canadian subjects (19 current depression.31 (0.72 (0.. retardation): 0.82 (0. Impact on quality of life measured with EQ-5D instrument.33) 0. anhedonia.13) 0.46 (0.58 (0.51 (0.33 (0.29) 0.27) 0. psychomotor agitation.284 Table 61: (Continued) Valuation method EQ-5D (TTO) 0. High-intensity psychological interventions ‘Remitters’: MADRS 12 ‘Responder’: at least 50% decrease in baseline score Schaffer et al.28) Psychological symptoms (low mood. 8 weeks’ follow-up.19) Study Definition of health states Sapin et al. guilt. noncomparative cohort study.74 (0.28) 0. from French primary care 8 weeks No depression Mild depression Moderate depression Severe depression Responder – remitter Responder – non-remitter Non-responders 0.34) 0.24) Somatic (decreased appetite.

2008b). which may limit their applicability to the UK setting. Full details of utility scores used in the model are presented in Table 62. ● Relapse: a linear increase from baseline utility to the ‘response – no remission’ health state (0. no relapse. the QALY profiles over 15 months were estimated when patients entered the three end health states in the model (no remission. ● Relapse: a linear increase from baseline utility to the ‘response with remission’ health state over the initial 3-month treatment period. and the valuation of health states be based on public preferences elicited using a choice-based method such as time trade-off (TTO) or SG. preference values assigned to health states were elicited from the UK population sample. NICE also suggests that the measurement of changes in HRQoL should be reported directly from people with the condition examined.15) over 15 months. ● No relapse: a linear increase from baseline utility to the ‘response with remission’ health state (0. 285 .15) to the ‘no response’ health state (0. For patients who did not complete their initial treatment. in a representative sample of the UK population (NICE. followed by linear decrease back to baseline utility over the remaining 12 months.85) over the initial 3-month treatment period. which is useful for modelling health states in cost-utility analyses. utility scores were stratified according to disease severity and clinical response. the following assumptions were used: ● No response: patient remains at baseline utility (0. remaining in the ‘response with remission’ health state for the following 12 months. Therefore.33 or 0.33 or 0. Although these utility scores were based on a cohort of French primary care patients.72) over 3 months. ● No relapse: a linear increase from baseline utility to the ‘response – no remission’ health state over 3 months. Furthermore. based on these recommendations. Estimation of quality-adjusted life years By applying the utility scores estimated by Sapin and colleagues (2004). decreasing immediately back to their baseline utility over the remaining 12 months. relapse) based on the following assumptions for patients who completed treatment: ● No remission: a linear increase from baseline utility score (0.High-intensity psychological interventions NICE currently recommends the EQ-5D as the preferred measure of HRQoL in adults for use in cost-utility analyses. The ICERs were calculated as the difference in the expected healthcare costs divided by the difference in the overall effectiveness of the two strategies. Incremental cost effectiveness of COMB versus antidepressant treatment The incremental cost effectiveness of COMB compared with antidepressant treatment for patients with moderate or severe depression was evaluated by assessing the difference in costs and effectiveness of each therapy. followed by linear increase to ‘response with remission’ health state over the remaining 12 months. the EQ-5D utility scores estimated by Sapin and colleagues (2004) were considered to be the most suitable for calculating QALYs in the guideline economic models. followed by a linear deterioration back to baseline utility over the remaining 12 months.58) over the initial 3-month treatment period.

24) (0.03) Guideline meta-analysis Blackburn et al.08 to 0. (2004) Sensitivity analyses Deterministic sensitivity analysis Given the considerable uncertainty around some of the input parameters used in the base case model and ambiguity surrounding any policy implications of point estimates.29 to 0. were explored. as well as the cost implications of different levels of resource use involved in patient clinical management. (1986).65 to 0.50 to 0.10 to 0.85 0.55 (0. one-way sensitivity analysis was undertaken. 286 .83 to 0.37) (0. Uncertainty around the various transition probabilities and quality-of-life weights.65 to 1.72 0.70 0.33 0.67 – (0.66) Sapin et al.58 0.30 0.30) GDG expert opinion Murphy et al.22) (0. (1984) (0.01) Guideline meta-analysis Base case value (mean) Range (95% CI) Source 0.76 0.81 0.55 to 1. Murphy et al.38 to 1.68 Probability of spontaneous remission 0.High-intensity psychological interventions Table 62: Clinical effectiveness parameters applied in the economic model Parameter Clinical outcomes Absolute risk of not completing treatment: AD RR of not completing treatment: COMB Absolute risk of no remission following treatment: AD RR of no remission following treatment: COMB Absolute risk of relapse during follow-up: AD RR of relapse during follow-up: COMB 0. This involved varying a single parameter between its plausible minimum and maximum values while maintaining all remaining parameters in the model at their base case value.87) (0.79) (0.15 0. (1984) (0.20 for patients who drop out of initial treatment: BOTH Probability of relapse for patients who discontinue initial treatment and in remission: BOTH Quality-of-life weights Moderate depression Severe depression Response with remission Response without remission No response 0.

gamma distributions to cost estimates and beta distributions to utility estimates and absolute rates. 1. a standard error based on 30% of the mean estimate was applied to reflect any potential uncertainty around these estimates. Effectiveness and cost estimates were then recalculated 10. Quality-adjusted life years The decision model for patients with moderate depression resulted in an average of 0.9. 95% CI 0.21). 95% CI 0.55.53 QALYs per patient in the COMB therapy group and 0. The two follow-up studies suggested that there is a lower risk of relapse in the COMB therapy arm (RR 0.000 times using Monte Carlo simulation. 2000).6 Results Clinical outcomes The systematic review of the clinical evidence showed that the probability of not completing the initial 3-month therapy was higher for AD than for COMB (RR 0. 8. 1. The decision model for patients with severe depression resulted in an average of 0. The expected subsequent health and social care cost over 15 months for patients who did 287 .76.67 QALYs per patient in the COMB therapy group and 0. plus 6 months’ maintenance therapy was £1314 for patients with moderate depression and £1643 for patients with severe depression.42 QALYs in the AD group. The expected subsequent health and social care cost over 15 months for patients who did not complete their initial therapy was £1638 for both moderate and severe patients.80. including a full course of CBT.11 per patient with severe depression.03).24) over a 12-month follow-up period although this was not statistically significant (p 0.01) while the probability of not achieving remission following therapy was also lower in the COMB group (RR 0. the average gain in QALYs over 15 months for COMB therapy was 0. 1.58 QALYs per patient in the AD group. Using the mean point estimates and their 95% CIs. Costs and cost effectiveness The full cost of a 3-month course of antidepressant treatment plus 6-month maintenance therapy was £270 for patients with moderate depression and £599 for patients with severe depression. appropriate distributions were assigned for each parameter estimate. Whether an intervention is cost effective or not depends on how decision makers value the additional health gain achieved by the therapy.High-intensity psychological interventions Probabilistic sensitivity analysis To demonstrate the joint uncertainty between the different parameters.68.09 per patient with moderate depression and 0. The probability that COMB therapy is cost effective compared with AD treatment as a function of decision makers’ maximum willingness-to-pay for an additional successfully treated patient or QALY was illustrated by CEACs (Briggs.38. probabilistic analysis is required. The full cost of 3-month COMB therapy.65. Therefore. For example. 95% CI 0. lognormal distributions were applied to relative risk estimates. For cost estimates that did not have 95% CIs.

High-intensity psychological interventions not respond to therapy and achieve remission was £1404 for both patient groups.7 Discussion In this economic evaluation.9. The resulting base case ICERs were £7. Sensitivity analyses Deterministic sensitivity analysis The parameter values used in the sensitivity analyses and the resultant ICERs are presented in Table 63. 8. An updated cost-effectiveness model was constructed to investigate the difference in clinical outcomes and direct health and social care costs between the different strategies. Incremental cost effectiveness of COMB versus antidepressant treatment Overall.052 per QALY gained for moderate depression and £5. The threshold value represents the maximum a decision maker would be willing to pay for a unit of effect. 2008a). This is explained by the high uncertainty around the relative risk estimate of no remission and to a lesser extent around the relative risk of relapse for COMB versus AD. The CEAC indicates the probability of COMB therapy being cost effective for a range of threshold values. these strategies were compared in a formal cost-effectiveness analysis. Other factors had a much lesser role in the variation of the results. The results of the deterministic sensitivity analysis indicated that the results were fairly robust when single parameters were varied over their uncertainty ranges. the strategy of COMB therapy was £624 more costly per patient with moderate depression and £653 more costly per patient with severe depression.000 to £30. COMB therapy was estimated to be significantly more effective and more costly than antidepressant treatment for patients with both moderate and severe depression. The cost-effectiveness estimates were most sensitive to: (1) the relative risk of no remission following therapy completion and.558 per QALY gained for severe depression. The updated clinical evidence review indicated that CBT alone may be more costly yet less clinically effective than antidepressant treatment and so it was excluded from the final model. CEACs were constructed (see Figure 7). Within this threshold range. As combination therapy is both more effective and more costly than antidepressant treatment. 288 . The expected subsequent health and social care cost of relapse while in remission was £702 for both patient groups. the probability of COMB therapy being cost effective for patients with moderate depression was 86 to 90% and for patients with severe depression was 88 to 92%. On average. in this case a QALY. Current NICE guidance sets a threshold range of £20.000 per QALY (NICE. (2) the relative risk of relapse while in remission. Probabilistic sensitivity analysis In order to present the results of the probabilistic sensitivity analysis. CBT was chosen as the psychological therapy and citalopram as the antidepressant drug being compared.

623 to 227.964 to 7.688 Relative risk of discontinuation Relative risk of non-remission Relative risk of relapse Relative risk of non-remission following discontinuation Quality-of-life weights 0.012 to 7.611 Moderate depression Severe depression Remission – no relapse Remission – relapse No Remission Resource use and costs 5 to 50% 8 to 16 N/A 2.391 to 5.000 5.773 to 7.143 to 11.222 6.092 6.535 to 5.65 to 1.573 to 7.254 5.141 N/A 5.08 to 0.762 to 7.385 4.55 to 1.22 0.558 Analysis Base case analysis Clinical efficacy (COMB versus AD) 0.606 N/A 6.205 5.03 0.187 to 5.79 0.50 to 0.052 Moderate depression Severe depression 5.984 to 5.866 2.420 3.87 0.148 5.610 to 7.65 to 0.83 to 0.355 7.9 6.558 N/A Severe patients – % receiving specialist care Moderate patients – number of CBT sessions High-intensity psychological interventions 289 .Table 63: Deterministic sensitivity analysis ICER per QALY (£) Uncertainty range 7.248 to 14.736 5.506 to 5.29 to 0.154 to 367.071 to 6.582 5.052 4.7 to 0.905 to 7.01 0.37 0.66 6.434 to 5.38 to 1.24 0.

.8 0. Limitations of the analysis The clinical effectiveness estimates used in the analyses were based on efficacy data obtained from RCTs. The difference in costs between combination therapy and antidepressant treatment was slightly higher for patients with severe depression. with remission rates applied at the end of the initial 3 months of treatment and relapse rates applied during the 12-month follow-up period. However.7 0.2 0. 2004).4 0. Another issue concerns the time horizon used for the analysis. One study in the clinical evidence review indicated lower relapse rates with combination therapy versus antidepressant treatment for up to 6 years after treatment (Fava et al. these costs were partially offset by savings due to lower subsequent treatment costs. A 15-month time horizon was used. 290 . resulting in possible over estimates of successful outcomes for both treatment options provided within the NHS setting. Overall.High-intensity psychological interventions Figure 7: CEACs of COMB therapy versus AD for patients with moderate and severe depression 1 Probability of COMB being cost effective 0.9 0. this is unlikely to significantly influence the relative effectiveness of the two treatment options. This suggests that the relative cost effectiveness of combination therapy versus antidepressant treatment may be underestimated when based on a short time horizon.5 0. while the difference in QALY gains was also slightly higher.1 0 0 5000 10000 15000 20000 25000 Cost-effectiveness Threshold per QALY (£) 30000 Moderate Depression Severe Depression Two separate analyses were conducted for patients with moderate and severe depression.3 0.6 0. The cost results for patients with both moderate and severe depression suggest that although the initial treatment cost of combination therapy is substantially higher. the results of the analysis indicate that combination therapy is likely to be a cost-effective first-line treatment for both moderate and severe depression.

This is because some key recommendations about psychological therapies are common to all types of interventions and also because a number of the recommendations draw on evidence from several different reviews. cognitive behavioural therapies have the largest evidence base. but as potentially the most cost-effective option because of its greater benefit despite the increased cost. and this is the focus of this section.High-intensity psychological interventions It would have been preferable to evaluate the two strategies over a longer follow-up period.1 Cognitive behavioural therapies With 46 studies. direct costs to patients and their families and lost productivity costs because of morbidity or premature mortality (Thomas & Morris. As this analysis was conducted from the health service and PSS perspective. However. but the lack of direct clinical evidence beyond 15 months precluded this.10. such as social service costs. Within this group of studies.10 FROM EVIDENCE TO RECOMMENDATIONS This section synthesises the evidence from the clinical summaries of all the psychological interventions reviewed in this chapter and the health economic evidence. the evidence indicates that psychological interventions have a beneficial effect in the treatment of people with depression and they do not have an increased risk for discontinuation when compared with antidepressants. McCrone et al. when determining treatment choice. It is likely that the inclusion of these costs would have further increased the probability of combination therapy being cost effective compared with antidepressant treatment. such non-healthcare costs were not considered. The GDG took the view that for patients with moderate depression a number of options. 8. 2008). The outcome of the model does not support the simple adoption of combination treatment as the first choice. which suggested that combination treatment is cost effective not just for severe depression but also for moderate depression.. as per NICE guidance. the evidence suggests that there are differences in the evidence base for the effectiveness among the psychological interventions reviewed in this chapter. The clinical effectiveness data also points to a clear advantage of combination treatment over antidepressants alone. 291 . 2003. This should then allow for a discussion between patient and clinician in which a number of factors are taken into account. Depression incurs significant non-healthcare costs. 8. Overall. This is supported by the outcome of the health economics model. including antidepressants alone and CBT alone (CBT alone was found to be better than antidepressants alone when both were compared with combined treatment). including the demands of adhering to the various treatment options and experience of past treatment. and as a result the recommendations from the previous guideline were changed. the largest dataset is that which compares individual CBT with antidepressants and which shows broad equivalence of effect across the range of severity. should be available.

the treatments set out in the trials. 2004a) and of their experiences of providing or receiving psychological treatments. broadly similar effects to CBT and. the GDG found evidence for some other interventions including IPT and couples therapy that suggested. the detailed outcomes of these reviews are set out in the sections for these interventions below. as far as possible. the GDG looked closely at the evidence for relapse. In making recommendations on CBT. There was no clear evidence for the effectiveness of other models of group CBT. The GDG also took into account the evidence on the delivery of effective psychological interventions reviewed in Chapter 6 and used this to develop a number of recommendations on the need to adhere to. to remove the previous recommendation about the provision of brief CBT because the GDG did not think that the rather limited evidence for brief CBT justified such a specific recommendation. for behavioural activation. Given the relapsing and remitting nature of depression. There was concern that this recommendation had led to an unnecessary restriction on the number of sessions of psychological intervention being made available. Instead. The provision of individual CBT is therefore one option when there are concerns over the risk of relapse (an almost everpresent concern with people who have had more than two episodes of depression) and should be considered along with the evidence reviewed for pharmacological interventions and relapse prevention (see Chapter 12). to a lesser extent. It also led the GDG. should that be achieved after fewer than the recommended sessions. and the data from psychological interventions specifically designed to reduce relapse. The GDG also reviewed the relative effectiveness of CBT against a range of other psychological interventions. the GDG were conscious of feedback from stakeholders from the previous guideline (NICE. as well as the need for routine outcome monitoring and the use of appropriate training and supervision.) 292 . In brief. The group CBT approach (based on the Coping with Depression model) showed evidence of benefit at post-treatment and at follow-up over a waitlist control. after considering the evidence. The most important evidence came from two sources: the studies comparing CBT with antidepressants. then treatment need not be continued beyond that point. The GDG did not consider the evidence for counselling. in some comparisons.High-intensity psychological interventions A more limited (because the evidence relates primarily to mild depression) dataset was examined for group CBT. which showed a reduced relapse rate for CBT in the follow-up of individual trials. Of the treatments specifically designed to reduce relapse. This led the group to specify in greater detail than previously the ways in which all psychological treatments in this guideline (including CBT) should be provided. (Note that this approach has been adopted for all recommendations for psychological therapies in this guideline. the GDG elected to recommend that the duration of treatment should be in line with that found in the majority of trials but also suggested that the target in treatment should be remission and. short-term psychodynamic psychotherapy or problem solving therapy to be as strong as that for other interventions (see below). group-based mindfulness-based cognitive therapy has the strongest evidence base with evidence that it is likely to be effective in people who have experienced three or more depressive episodes.

It is also a component part of cognitive behavioural interventions for depression and one of the first important trials of behavioural activation was a deconstruction study (Jacobson1999). it could be considered as an option. 8. having considered the evidence on clinical and cost effectiveness.High-intensity psychological interventions The GDG felt it was important to locate all psychological interventions in proper relation to each other. CBT and usual care. however. The GDG decided that although the evidence was not sufficiently robust to recommend behavioural activation as a direct alternative individual treatment option to CBT or IPT. given the duration of the group and the staffing of such groups. consideration was given to the emergence of new evidence. Group CBT is an effective treatment for mild depression but. all of which were positive for behavioural activation (that is. the GDG decided not to 293 .10. The GDG did. However.2 Behavioural activation There has been renewed interest in behavioural activation as a treatment for depression and a number of new studies were identified for the review in this guideline. including data on comparisons with placebo. there was no evidence of the superiority of these other interventions). and therefore are to be preferred over individual and group CBT (and other high-intensity psychological interventions) as the initial treatment for subthreshold depressive symptoms and mild depression. antidepressants. it was viewed on cost-minimisation grounds as less cost effective than low-intensity interventions but more cost effective than individual CBT. decide that healthcare professionals should be made aware of the more limited evidence base for behavioural activation when compared with CBT. It was also not possible to evaluate the cost effectiveness of behavioural activation in an economic model because of the limited clinical evidence available. This meant developing recommendations that locate all psychological interventions at the appropriate place in the stepped-care model.3 Problem solving therapy Problem solving therapy was recommended as a separate intervention in the last guideline. 8. In light of the improved evidence for a range of low-intensity interventions that have emerged since the last guideline. No direct evidence on the cost effectiveness of behavioural activation as a high. Low-intensity interventions are clinically and cost effective for subthreshold depressive symptoms and mild depression. leaving a limited dataset based only on two studies with much of the evidence for effectiveness being dependent on one study (Mynors-Wallis1995). Note was also taken of the evidence for the effectiveness of behavioural activation in low-intensity interventions. No new studies were identified.10.or low-intensity intervention was identified in the systematic review of the health economics literature. and so was placed between them in the stepped-care model. IPT and couples therapy (see below). although it should be noted that the duration and frequency of high-intensity behavioural activation is identical to that of high-intensity CBT.

couples therapy (based on a behavioural model) is recommended. but for many patients with mild to moderate depression IPT is an appropriate alternative to CBT. depression and life difficulties (Kendrick et al. the use of combination CBT and antidepressant drugs as the initial treatment for severe depression). 8.. As a result of the increased evidence identified in this guideline. There was also no good economic evidence for the effectiveness of IPT and it was also not possible to evaluate the cost effectiveness of IPT in an economic model because of the limited clinical evidence available. 8.or individual-based approaches to relapse prevention). but rather decided that it should be focused on patients in established relationships where the relationship may play a role in the development. 2000). those who adopted a behavioural approach to treatment) when compared with waitlist control. because these issues were typical of the patients who entered the trials reviewed in this guideline. 294 .. nor is the evidence for the range of applications for IPT as strong or as wide ranging (for example the evidence on group.High-intensity psychological interventions recommend problem solving therapy as a separate intervention in this guideline. one study was found that suggested that problem solving therapy delivered by community mental health nurses was not cost effective compared with usual GP care in patients with new episodes of anxiety. which suggested that couples therapy may be a cost-effective treatment compared with antidepressant medication for patients with major depressive disorder (Leff et al. The evidence base for couples therapy is relatively modest with just six studies. the GDG did not develop recommendations for IPT that were as broad in scope as for CBT (for example. the dataset is not as large as that for CBT. In the health economics literature. Only one study was identified in the health economics literature. 2006a).10. however. but there are indications of a beneficial effect in couples with depression (in particular. a number of additional studies were included and one from the previous guideline was excluded. but this study was excluded from the clinical evidence analyses because of its high attrition rate.5 Interpersonal therapy The evidence for the effectiveness of IPT reviewed in this guideline confirms the picture in the previous guideline of IPT as an effective treatment for depression. However. the GDG did not consider it appropriate to offer it as a direct alternative to CBT or IPT. maintenance or resolution of depression.10. However.4 Couples therapy In the review for this guideline. and evidence of similar outcomes for couples therapy when compared with individual CBT and IPT (although the evidence in comparison with IPT is more uncertain). 2008). Therefore. the GDG expects that it will continue to be one of the component parts of the low-intensity interventions offered for the treatment of depression (Richards & Whyte.

.6 Counselling The evidence base for counselling identified in the previous guideline was small (only three studies – one of which. Nevertheless. Stiles et al. while one did not meet inclusion criteria and two other studies compared different forms of counselling (GOLDMAN2006.High-intensity psychological interventions 8. GREENBERG1998). Two studies identified in the health economics literature suggested no advantage. IPT. taking the total number of included studies to ten. An inconsistent picture of the effectiveness of counselling emerges from the review. with one trial having poorer outcomes against usual care (Simpson2003) and one against antidepressants (Bedi2000). SALMINEN2008) and some of the largest effects were obtained in dysthymic or subthreshold populations (MAINA2005). of counselling compared with either usual GP care or antidepressant treatment in adults with depression (Friedli et al. 2000. a review of the practice-based evidence did not provide clear evidence of a benefit for counselling in depression (for example. Two studies identified in the health economics literature suggested no advantage. Ward2000.. Miller et al. such as CBT. only one provided relevant data on an important comparison relevant to the effectiveness of counselling (WATSON2003). from a review of these studies it was not possible to demonstrate a consistent picture of any clinically important benefit for short-term psychodynamic psychotherapy in depression. The comparators were very varied and so significantly limited the amount of metaanalysis that was possible. the GDG decided not to support the same recommendation for counselling in this guideline update. Marriott & Kellett. 2003). Nevertheless.7 Short-term psychodynamic psychotherapy A number of new studies were identified for short-term psychodynamic psychotherapy as a treatment for depression. 2008. did not meet current inclusion criteria). The previous guideline recommended counselling in mild to moderate depression. Of the new studies identified. Furthermore. 8. CBT. IPT. but felt that the limited evidence should be drawn to the attention of the healthcare professional. but in light of the increased evidence for a range of lowintensity interventions and group CBT for mild to moderate depression.10. 2009). in terms of cost effectiveness. of short-term psychodynamic psychotherapy compared with usual care for primary care patients with depression (Simpson et al. 2003) or high utilisers of psychiatric services 295 . in terms of cost effectiveness. the GDG thought that counselling may be considered for people with mild to moderate depression who have declined an antidepressant.. WATSON2003). but no difference identifiable in the two comparisons with CBT (Ward2000. There was considerable discussion of this recommendation in the GDG. the two comparisons with antidepressants revealed directly contradictory results (DEKKER2008. which took into account not only the limited evidence for counselling but also the increased evidence for other interventions.10. The evidence base for counselling in contrast to that for both CBT and IPT lacked data on both long-term follow-up and relapse prevention. behavioural activation and behavioural couples therapy.. behavioural activation or behavioural couples therapy. For example.

1 8. The previous guideline recommended psychodynamic psychotherapy for complex comorbidities. Practitioners should consider using competence frameworks developed from the relevant treatment manual(s) and for all interventions should: ● receive regular high-quality supervision ● use routine outcome measures and ensure that the person with depression is involved in reviewing the efficacy of the treatment ● engage in monitoring and evaluation of treatment adherence and practitioner competence . behavioural activation or behavioural couples therapy.1. behavioural activation and behavioural couples therapy.1).11. There was considerable discussion of this recommendation in the GDG which took into account not only the limited evidence for short-term psychodynamic psychotherapy but also the increased evidence for other interventions such as CBT. As a result of the limited evidence for short-term psychodynamic psychotherapy for depression with or without complex comorbidities.11. IPT. and external audit and scrutiny where appropriate.11 8. Psychological and psychosocial interventions should be based on the relevant treatment manual(s).11.2. which included a subgroup analysis of those with a personality disorder.5. but that the limited evidence should be drawn to the attention of the healthcare professional. by using video and audio tapes. Nevertheless. IPT.2 8.1 RECOMMENDATIONS Effective delivery of interventions for depression All interventions for depression should be delivered by competent practitioners. 8.1 Group cognitive behavioural therapy Consider group-based CBT for people with persistent subthreshold depressive symptoms or mild to moderate depression who decline low-intensity psychosocial interventions (see 7. 1999). CBT. 296 . 8. the short-term psychodynamic psychotherapy evidence base lacked data on both long-term follow-up and relapse prevention. lacked the power to inform a decision on the use of short-term psychodynamic psychotherapy with comorbidities such as personality disorder. which should guide the structure and duration of the intervention. the GDG took the view that short-term psychodynamic psychotherapy may be considered for people with mild to moderate depression who have declined an antidepressant.1.. the GDG did not feel able to endorse the recommendation in the previous guideline and developed a more specific recommendation for this update.11.High-intensity psychological interventions with a significant number of patients with a diagnosis of depression (Guthrie et al. but the current dataset offered no clear evidence for the effectiveness of short-term psychodynamic psychotherapy for complex comorbidities.for example. As with the evidence base for counselling. Results from the KOOL2003 study.

297 . behavioural activation and behavioural couples therapy.3.2 8. CBT.11. IPT. The choice of intervention should be influenced by the: ● duration of the episode of depression and the trajectory of symptoms ● previous course of depression and response to treatment ● likelihood of adherence to treatment and any potential adverse effects ● person’s treatment preference and priorities72. 8.11.interpersonal therapy (IPT) . For people with depression who decline an antidepressant. consider: ● counselling for people with persistent subthreshold depressive symptoms or mild to moderate depression ● short-term psychodynamic psychotherapy for people with mild to moderate depression.2 Group-based CBT for people with persistent subthreshold depressive symptoms or mild to moderate depression should: ● be based on a structured model such as ‘Coping with Depression’ ● be delivered by two trained and competent practitioners ● consist of ten to 12 meetings of eight to ten participants ● normally take place over 12 to 16 weeks.11. or where involving the partner is considered to be of potential therapeutic benefit70.3 8.behavioural couples therapy for people who have a regular partner and where the relationship may contribute to the development or maintenance of depression.4 70This 71Ibid. normally one of the following options: . including follow-up. 72Ibid.High-intensity psychological interventions 8.11.3 8. 8. discuss the relative merits of different interventions with the person and provide: ● an antidepressant (normally a selective serotonin reuptake inhibitor [SSRI]) or ● a high-intensity psychological intervention.3. For people with moderate or severe depression.behavioural activation (but note that the evidence is less robust than for CBT or IPT) .2. recommendation also appears in the chapter on pharmacological interventions.3.11.11.1 Treatment options For people with persistent subthreshold depressive symptoms or mild to moderate depression who have not benefited from a low-intensity psychosocial intervention.3. provide a combination of antidepressant medication and a high-intensity psychological intervention (CBT or IPT)71.CBT .

4.11. 8.2 8.11. As the aim of treatment is to obtain significant improvement or remission the duration of treatment may be: ● reduced if remission has been achieved ● increased if progress is being made. For people with severe depression.5 8.4. For all people with depression having IPT. 298 .11. if there is a comorbid personality disorder or significant psychosocial factors that impact on the person’s ability to benefit from treatment). Also consider providing: ● two sessions per week for the first 2 to 3 weeks of treatment for people with moderate or severe depression ● follow-up sessions typically consisting of three to four sessions over the following 3 to 6 months for all people with depression. Also consider providing: ● two sessions per week for the first 3 to 4 weeks of treatment for people with moderate or severe depression ● follow-up sessions typically consisting of three to four sessions over the following 3 to 6 months for all people with depression.1 Delivering high-intensity psychological interventions For all high-intensity psychological interventions. For all people with depression having behavioural activation. the duration of treatment should normally be within the limits indicated in this guideline.5 Delivering counselling 8. the duration of treatment should typically be in the range of 16 to 20 sessions over 3 to 4 months.4.High-intensity psychological interventions Discuss with the person the uncertainty of the effectiveness of counselling and psychodynamic psychotherapy in treating depression.1 For all people with persistent subthreshold depressive symptoms or mild to moderate depression having counselling. Behavioural couples therapy for depression should normally be based on behavioural principles.11.11.11.4. For all people with depression having individual CBT.3 8. 8. the duration of treatment should typically be in the range of 16 to 20 sessions over 3 to 4 months. and an adequate course of therapy should be 15 to 20 sessions over 5 to 6 months. the duration of treatment should typically be in the range of six to ten sessions over 8 to 12 weeks. and there is agreement between the practitioner and the person with depression that further sessions would be beneficial (for example.4 8.11. the duration of treatment should typically be in the range of 16 to 20 sessions over 3 to 4 months.4.4 8. consider providing two sessions per week for the first 2 to 3 weeks of treatment.11.5.

11. consider combining antidepressant medication with CBT73.High-intensity psychological interventions 8. the duration of treatment should typically be in the range of 16 to 20 sessions over 3 to 4 months.9.7 8. 8.1 Combined psychological and drug treatment For a person whose depression has not responded to either pharmacological or psychological interventions.11.11.9.6.2 73This recommendation can also be found in the chapter on pharmacological interventions.11. should be offered the following psychological interventions: ● individual CBT for people who have relapsed despite antidepressant medication and for people with a significant history of depression and residual symptoms despite treatment ● mindfulness-based cognitive therapy for people who are currently well but have experienced three or more previous episodes of depression.1 Delivering short-term psychodynamic psychotherapy For all people with mild to moderate depression having short-term psychodynamic psychotherapy. the duration of treatment should typically be in the range of 16 to 20 sessions over 4 to 6 months.6 8.9 8.1 Delivering psychological interventions for relapse prevention For all people with depression who are having individual CBT for relapse prevention. Mindfulness-based cognitive therapy should normally be delivered in groups of eight to 15 participants and consist of weekly 2-hour meetings over 8 weeks and four follow-up sessions in the 12 months after the end of treatment. 299 .11. 8. If the duration of treatment needs to be extended to achieve remission it should: ● consist of two sessions per week for the first 2 to 3 weeks of treatment ● include additional follow-up sessions.7. 8.8.1 Psychological interventions for relapse prevention People with depression who are considered to be at significant risk of relapse (including those who have relapsed despite antidepressant treatment or who are unable or choose not to continue antidepressant treatment) or who have residual symptoms.11. typically consisting of four to six sessions over the following 6 months.8 8.11.11. 8.11.

However the practicality. The results of this study will have important implications for the provision of psychological treatment in the NHS. In well-defined depression of moderate to severe severity. 8. and mediators and moderators of response should be investigated. The availability of alternatives drawing from a different theoretical model is therefore important.1. This question should be answered using a randomised controlled trial design that reports short. acceptability and cost effectiveness of combined treatment over a sequenced approach is less well-established. CBT has the best evidence base for efficacy but it is not effective for everyone.and medium-term outcomes (including cost-effectiveness outcomes) of at least 18 months’ duration. The answer has important practical implications for service delivery and resource implications for the NHS. Psychotherapy based on psychodynamic principles has historically been provided in the NHS but provision is patchy and a good evidence base is lacking.1 RESEARCH RECOMMENDATIONS The efficacy of short-term psychodynamic psychotherapy compared with CBT and antidepressants in the treatment of moderate to severe depression.12. or single modality treatment with the addition of the other modality if there is inadequate response to initial treatment.1.and patient-rated assessments of improvement and an assessment of the acceptability of the treatment options. It is therefore important to establish whether short-term psychodynamic psychotherapy is an effective alternative to CBT and one that should be provided.2 The cost effectiveness of combined antidepressants and CBT compared with sequenced treatment for moderate to severe depression What is the cost effectiveness of combined antidepressants and CBT compared with sequenced medication followed by CBT and vice versa for moderate to severe depression? Why this is important There is a reasonable evidence base for the superior effectiveness of combined antidepressants and CBT over either treatment alone in moderate to severe depression. The outcomes chosen should reflect both observer. Particular attention should be paid to the reproducibility of the treatment model and training and supervision of those providing interventions to ensure that the treatments are both robust and generalisable. The outcomes chosen should reflect both observer and patient-rated assessments for acute and medium-term 300 . what is the efficacy of short-term psychodynamic psychotherapy compared with CBT and antidepressants? Why this is important Psychological treatments are an important therapeutic option for people with depression.12 8. The study needs to be large enough to determine the presence or absence of clinically important effects using a noninferiority design.12.High-intensity psychological interventions 8. This question should be answered using a randomised controlled trial design in which people with moderate to severe depression receive either combined treatment from the outset.

The study needs to be large enough to determine the presence or absence of clinically important effects using a non-inferiority design together with robust health economic measures. 8. what is the efficacy of counselling compared with low-intensity cognitive behavioural interventions? Why this is important Psychological treatments are an important therapeutic option for people with subthreshold symptoms and mild depression. and an assessment of the acceptability of the treatment options.High-intensity psychological interventions outcomes to at least 6 months. but the best way to treat them is not known.12. 8. as are the outcomes for counselling.4 The efficacy of counselling compared with low-intensity cognitive behavioural interventions and treatment as usual in the treatment of persistent subthreshold depressive symptoms and mild depression In persistent subthreshold depressive symptoms and mild depression. A careful definition of persistence should be used which needs to include duration of symptoms and consideration of failure of low-intensity interventions and does not necessarily imply a full diagnosis of dysthymia.1.12. The results of this study will have important implications for the provision of psychological treatment in the NHS. There are studies of the efficacy of antidepressants for dysthymia (persistent subthreshold depressive symptoms that have lasted for at least 2 years) but there is a lack of evidence for CBT. The outcomes chosen should reflect both observer and patient-rated assessments of improvement. This research recommendation is aimed at informing the treatment options available for this group of people with subthreshold depressive symptoms that persist despite low-intensity interventions.3 The efficacy of CBT compared with antidepressants and placebo for persistent subthreshold depressive symptoms What is the efficacy of CBT compared with antidepressants and placebo for persistent subthreshold depressive symptoms? Why this is important Persistent subthreshold depressive symptoms are increasingly recognised as affecting a considerable number of people and causing significant suffering. The study needs to be large enough to determine the presence or absence of clinically important effects using a non-inferiority design. Subthreshold depressive symptoms of recent onset tend to improve but how long practitioners should wait before offering medication or psychological treatment is not known. It is therefore important to establish whether either of these interventions is an effective alternative to treatment as usual and should be provided in the NHS.and medium-term outcomes (including cost-effectiveness outcomes) of at least 6 months’ duration. and an assessment of the acceptability and burden of the treatment options. 301 .1. This question should be answered using a randomised controlled trial design that reports short. and mediators and moderators of response should be investigated. Low-intensity cognitive behavioural interventions have the best evidence base for efficacy but the evidence is limited and longer-term outcomes are uncertain.

The results of this study will have important implications for the provision of psychological treatment in the NHS. It is therefore important to establish whether behavioural activation is an effective alternative to CBT and one that should be provided. 8.5 The efficacy of behavioural activation compared with CBT and antidepressants in the treatment of moderate to severe depression In well-defined depression of moderate to severe severity. The outcomes chosen should reflect both observer and patient-rated assessments of improvement and an assessment of the acceptability of the treatment options. This question should be answered using a randomised controlled trial design which reports short-term and medium-term outcomes (including cost-effectiveness outcomes) of at least 18 months’ duration. Particular attention should be paid to the reproducibility of the treatment model and training and supervision of those providing interventions in order to ensure that the treatments are both robust and generalisable.6 The efficacy and cost effectiveness of cognitive behavioural therapy. interpersonal therapy and antidepressants in prevention of relapse in people with moderate to severe recurrent depression In people with moderate to severe recurrent depression. what is the efficacy of behavioural activation compared with CBT and antidepressants? Why this is important Psychological treatments are an important therapeutic option for people with depression.1. Particular attention should be paid to the reproducibility of the treatment model and training and supervision of those providing interventions in order to ensure that the treatments are both robust and generalisable. The availability of alternatives drawing from a different theoretical model is important because outcomes are modest even with the best supported treatments. The outcomes chosen should reflect both observer and patient-rated assessments of improvement.12. The study needs to be large enough to determine the presence or absence of clinically important effects using a non-inferiority design.High-intensity psychological interventions This question should be answered using a randomised controlled trial design which reports short-term and medium-term outcomes (including cost-effectiveness outcomes) of at least 18 months’ duration. The study needs to be large enough to determine the presence or absence of clinically important effects using a non-inferiority design.12. and an assessment of the acceptability of the treatment options.1. 8. what is the relative efficacy of CBT. and mediators and moderators of response should be investigated. but evidence on the prevention of relapse (especially for 302 . and mediators and moderators of response should be investigated. IPT and antidepressants in preventing relapse? Why this is important Psychological and pharmacological treatments are important therapeutic options for people with depression. Behavioural activation is a promising treatment but does not have the substantial evidence base that CBT has.

The study needs to be large enough to determine the presence or absence of clinically important effects using a non-inferiority design. New developments in the style and delivery of CBT and IPT show some promise in reducing relapse but need to be tested in a large-scale trial. including the nature and duration of the intervention. All of these treatments have shown promise in reducing relapse but the relapse rate remains high. The results of this study will have important implications for the provision of psychological treatment in the NHS. and mediators (including the focus of the interventions) and moderators (including the severity of the depression) of response should be investigated. 303 .High-intensity psychological interventions psychological interventions) is limited. This question should be answered using a randomised controlled trial design which reports short-term and medium-term outcomes (including cost-effectiveness outcomes) of at least 24 months’ duration. The outcomes chosen should reflect both observer and patient-rated assessments of improvement and an assessment of the acceptability of the treatment options. Particular attention should be paid to the development and evaluation of CBT. IPT and medication interventions tailored specifically to prevent relapse.

relapse prevention and next-step treatments (treatments for treatment-resistant depression in the previous guideline). Kirsch et al.. 2008). discontinuation. new narrative reviews of transcranial magnetic stimulation (TMS) and vagus nerve stimulation (VNS). It discusses some of the issues that the GDG addressed in assessing the evidence base in order to form recommendations. duloxetine and therapies for depression with a seasonal pattern. Since the introduction of the MAOIs and the first TCA. 2002b). Excellent reviews of the topic are to be found in the British Association for Psychopharmacology Evidence-Based Guidelines for Treatment of Depressive Disorder (Anderson et al.. and new sections for chronic depression and residual symptoms. Of this patient group.Introduction to pharmacological and physical interventions 9 INTRODUCTION TO PHARMACOLOGICAL AND PHYSICAL INTERVENTIONS 9. 2008. dosage. approximately 20% will respond with no treatment at all. including that of placebo response. 30% will respond to placebo and 50% will respond to antidepressant drug 74Recommendations from TA59 were incorporated into the previous depression guideline according to NICE protocol. imipramine. 2002a. the more severe the symptoms. There are also new reviews for TCAs. In general. The severity of depression at which antidepressants show consistent benefits over placebo is poorly defined.. 2003)74.. The reviews of pharmacological interventions themselves are presented in the following chapters. 304 . cardiotoxicity. There has been intensive research on the effects of drug therapy on depression and how drugs might alter the natural history of the disorder. 2008) and in the World Federation of Societies of Biological Psychiatry’s (WFSBP) Guidelines for the Biological Treatment of Unipolar Depressive Disorders Parts 1 and 2 (Bauer et al. the greater the benefit (Anderson et al. and the following narrative reviews have been updated with new data: effect of sex on antidepressant choice. Where reviews have not been updated. The scope for the update also includes updating the NICE technology appraisal on the use of ECT (for depression) (TA59. in the late 1950s. antidepressants are normally recommended as first-line treatment in patients whose depression is of at least moderate severity.1 INTRODUCTION For the guideline update the following reviews of pharmacological interventions are updated: escitalopram. and antidepressants and suicide. an explanation has been added to the relevant chapter introduction. NICE. This chapter introduces the pharmacological interventions in the management of depression covered by this guideline update (although other physical interventions are also reviewed). many new antidepressants have been introduced and approximately 35 different antidepressants in a number of classes are currently available worldwide.

2008. a somewhat arbitrary dichotomy. the most effective way to prevent suicidal thoughts and acts is to treat depression. as have drugs with a similar pharmacology that are used for an indication other than depression (for example.. switching to another antidepressant. such as nausea.and long-term treatment. Most side effects of antidepressants are dose related. Anderson et al. which limit adherence to treatment with antidepressants. SSRIs as a class are associated with headache and gastrointestinal symptoms. Overall. 1987). These include dose escalation. are more effective than placebo in the treatment of major depression. atomoxetine). NICE. and a relative higher propensity than other antidepressants to cause sexual dysfunction.. 2005. tend to persist. and are generally equally effective (Cochrane Database of Systematic Reviews. tachycardia and corrected QT interval (QTc) prolongation.. It has been proposed that early non-persistent improvement in depressive symptoms may be due to a placebo response (Quitkin et al. sexual dysfunction. hyponatraemia and gastrointestinal bleeds. particularly in adolescents and young adults. measured using continuous scales. Some common antidepressant side effects.Introduction to pharmacological and physical interventions treatment (Anderson et al. tends to show a relatively smaller mean difference between active treatment and placebo. such as anticholinergic effects and. in some patients. 2008). when considered individually or by class. however. that response in clinical trials is generally defined as a 50% reduction in depression rating scale scores. are generally better tolerated than antidepressants from other classes and most are available as generic preparations.. There are concerns over side effects following short. but not as well tolerated as SSRIs. and that change. TCAs tend to be associated with a high burden of anticholinergic side effects and a higher propensity than other antidepressants to cause adverse cardiovascular effects including hypotension. Antidepressant treatment has been associated with an increased risk of suicidal thoughts and acts.2). and SSRIs are now the most commonly prescribed group of antidepressants in the UK (see also Section 9. Overall. Although the relative risk of developing suicidal thoughts and acts may be elevated above placebo rates in some patient groups. At the present time there are a variety of strategies for improving efficacy following initial non-response that are supported by existing evidencebased guidelines or systematic reviews. and combining the antidepressant with another antidepressant. 2008). It should be noted. tend to resolve within the first week of treatment whereas others. the absolute risk remains very small. This gives a number needed to treat (NNT) of three for antidepressants over waitlist control and five for antidepressants over placebo. SSRIs are considerably safer in overdose than TCAs. Systematic reviews using meta-analysis suggest that antidepressant drugs. leading to the recommendation that patients should be warned of this potential adverse effect during the early weeks of treatment and know how to seek help if required. but recent evidence has emphasised that improvement starts immediately on commencing treatment and early improvement is a strong predictor of eventual response which is unlikely if no improvement is evident after 4 weeks of treatment (Posternak & Zimmerman. venlafaxine is better tolerated than TCAs. 305 . An SSRI was recommended as first-line pharmacological treatment of moderate to severe depression in the previous guideline. 2004a). All antidepressants have been implicated. Gartlehner et al.

In the 3 months to June 2008. Section 11.5 million for tricyclic and related antidepressants (over half of which were for amitriptyline).nhsbsa. Solomon et al. a second generation antipsychotic or thyroid hormones. An untreated depressive episode typically lasts about 6 months (Angst & Preisig. 1997) and. is also supported by an evidence base.. 306 . It is also recommended that patients with recurrent major depression should go on to receive maintenance antidepressant drug treatment (NICE.Introduction to pharmacological and physical interventions a second drug such as lithium. and over 1 million for other antidepressants (the vast majority of which were for venlafaxine or mirtazapine). Systematic assessment of the evidence for these strategies is a major feature of this guideline update. However. The number of prescriptions for antidepressants dispensed by community pharmacies in England has risen steadily over the last 15 years.uk).nhs. All antidepressant drugs can cause discontinuation symptoms with short half-life drugs being most problematic in this respect (see Chapter 11. During the launch year.9 million prescriptions for antidepressants were dispensed by community pharmacists in England at a total cost of £18. 9. the costs of individual drugs and prescribing trends can be found on the NHS Business Authority website (www. but not always. an epidemiological study conducted in 1995 found that treatment remained suboptimal (Lepine et al. over 4. and longer if there are factors that increase the risk of relapse. Details of the number of antidepressant prescriptions dispensed in primary care. for an antidepressant drug.5 million prescriptions were dispensed for SSRIs (almost half of which were for citalopram). It is recommended that the same dose of antidepressant is used in this continuation phase. 2004a). 1993).2 DOSE AND DURATION OF ANTIDEPRESSANT TREATMENT: EVIDENCE FROM CLINICAL PRACTICE Prevalence of antidepressant prescribing 9. Only a third of people with major depression in the UK received a prescription usually. 1997). the Royal College of Psychiatrists launched the ‘Defeat Depression’ campaign to raise public awareness of depression and improve treatment (Vize & Priest. it is currently recommended that antidepressant drug treatment is continued for a minimum of 6 months after remission of major depression (12 months in older adults).2. The recurrence rate is lower when treatment is maintained with the effective acute treatment dose compared with a reduction to half the dose. 9. in view of the high recurrence rate if antidepressant medication is stopped immediately after response. There is good evidence that patients with residual symptoms are at increasing risk of relapse of major depression and the current practice is to continue treatment for longer in those patients. 1995. Adjunctive use of psychological therapies. over 2.. costs are falling due to the availability of an increasing number of antidepressants as generic preparations.8). particularly CBT. Although the number of prescriptions written continues to increase.1 million.1 In 1992.

a systematic review of the efficacy and tolerability of low versus high doses of TCAs was undertaken. 1996). Sixty percent of the patients taking fluoxetine completed 6 months of treatment compared with less than 40% of the patients taking TCAs. There is some evidence that the mean figure quoted for SSRIs may mask important differences between drugs: Donoghue (2000) found that. Those who discontinued one antidepressant were offered another.8% if lofepramine was excluded) (Dunn et al. 2005)... The prescribing of TCAs in this way is known to be pervasive across different countries and over time (Donoghue. 1999). 1996). while there was insufficient evidence to determine if there was a difference with respect to response or continuous endpoint data.. SSRIs are consistently found to be prescribed ‘at an effective dose’ in a much greater proportion of cases than TCAs. a UK prescribing study collecting data from over 750.204 patients who were prescribed TCAs or SSRIs by their GP.150 patients who were prescribed SSRIs.1% of TCA prescriptions were for ‘an effective dose’ compared with 99. in a GP population of 6. For example. 33% of those prescribed an SSRI completed ‘an adequate period of treatment’ compared with 6% of those prescribed a TCA (2.2. A more recent. In the previous guideline.2.. It is generally accepted that response to TCAs is partially dose-related. An RCT conducted in the US randomised 536 adults to receive desipramine. but no such effect has been demonstrated for SSRIs.000 patient records found that if lofepramine was excluded the mean doses prescribed for individual TCAs fell between 58 mg and 80 mg. randomised UK study also found that there was a higher rate of switching to another antidepressant with TCAs (including lofepramine) than SSRIs (Peveler et al. Donoghue & Hylan. 27% of patients taking fluoxetine were still receiving prescriptions after 120 days compared with 23% of patients taking paroxetine and 13.2 Dose Studies of prescribing practice have generally taken 125 mg and above of TCAs (except lofepramine) and licensed doses of SSRIs to be ‘an effective dose’. Of course. A further UK study that followed prescribing for 20. and compared prescribing in practice with this ideal.3 Duration In a UK study of 16.9% of prescriptions for SSRIs (Donoghue et al. no difference was found with respect to remission data. 1996). There were no differences in overall completers or response rates at endpoint. 2000. 2001). imipramine or fluoxetine (Simon et al. Only 13.5% of patients taking sertraline.195 GP patients found that at least 72% of those prescribed TCAs never received ‘an effective dose’ compared with 8% of those prescribed SSRIs (MacDonald et al. suggesting that initial drug choice did not affect outcome. prescribing patterns cannot be directly linked with outcome in studies of this type. ‘An adequate period of treatment’ was defined by the authors as: prescriptions covering at least 120 days’ treatment within the first 6 months after diagnosis.. naturalistic. outside of 307 . However. 9.Introduction to pharmacological and physical interventions 9.

42% of patients taking an SSRI were still receiving prescriptions compared with 27% of patients taking a TCA. This constitutes a treatment in itself and almost 30% of patients assigned to placebo respond within 6 weeks (Walsh et al. In some trials the participants are allowed to contact the therapist at any time to report problems. either because they drop out of treatment themselves. In short. although at that time it often contained an active ingredient. or they are withdrawn from the RCT by the anxious clinician (for example. they receive everything except the pharmacological help from the tablet in the ‘active drug’ arm of the trial. patients are assigned randomly to different treatment arms to reduce bias and therefore to reduce systematic differences in the allocation of patients that might affect the results. Stassen et al.. There is some evidence from this study that the relapse rate may have been higher in the TCA group: 28% of TCA-treated patients received a subsequent prescription for an antidepressant after a 9-month treatment-free gap compared with 10% of patients taking an SSRI. The issue of placebo response is discussed further in section 9..Introduction to pharmacological and physical interventions clinical trials. A placebo is an inert or innocuous substance that began to be used increasingly in control conditions in clinical trials during the 1950s. This response can include spontanous improvement. 2002]) because clinicians are reluctant to allow suicidal patients.. 1999). 1993. the chance of improvement is greater.6. after 6 months. 1993). patients may not return to their GP to have their treatment changed and outcome may be less positive. Next. of those enrolled into an RCT.. 9. 308 . Worse still. Khan et al.. Unfortunately. with shorter and milder depression. or patients with severe degrees of depressive phenomena. The response of patients to the inert substances now used should not be equated with the untreated course of the disorder because patients taking placebo have regular meetings with their doctor and receive supportive help. results are often presented only for ‘completers’ rather than for the full ‘intention-to-treat’ sample. which is a function of the duration and severity of the disorder. High spontaneous improvement rates are a major cause of ‘failed trials’ where active treatment is not statistically significantly more effective than placebo. a Swedish study of 949 patients found that 35% only ever received one prescription irrespective of whether it was for a TCA or an SSRI (Isacsson et al. 2002). typically 20 to 35% fail to complete the study. there is a tendency for investigators to recruit patients with less severe depression to RCTs and these are more likely to recover spontaneously (Khan et al.3 LIMITATIONS OF THE LITERATURE: PROBLEMS WITH RANDOMISED CONTROLLED TRIALS IN PHARMACOLOGY In RCTs. 2002). Conversely. patients with more severe depression are less likely to be thought suitable for RCTs (despite being more likely to show a true drug effect [Angst. Primary efficacy is usually based on a placebo-controlled RCT in which one of the treatment arms is a ‘placebo’ treatment. to run the risk of being randomised to an inactive treatment. For example.

Thase. Most studies of the effects of drugs are sponsored by the drug industry and these have been shown to be more than four times as likely to demonstrate positive effects of the sponsor’s drug as independent studies (Lexchin et al. Although raters may be blinded to the treatment arm to which a patient is allocated. These all add ‘noise’ to the assessment leading to increased variability and make it difficult to assess the ‘true’ size of any treatment effect. which means that subsequent ratings from an extreme value (such as high depression score) will tend to drop simply by virtue of being remeasured.. Finally. the GDG has been careful to consider their application to routine practice.. However. which not only leads to patients with milder degrees of depression being studied but also may contribute to the drop in scores after the treatment has started when severity may be more realistically assessed. so that patient and rater expectations of improvement may confound assessments. 2009). Therefore the bulk of the guideline recommendations are based on RCT evidence. In the recent naturalistic STAR*D study. 309 . only 22% of depressed patients met typical criteria for a phase III clinical trial. introduces a severe bias in the other direction. taken together with the smaller likelihood of severely depressed patients being included. while intention-to-treat analysis... cit. there is the phenomenon of ‘regression to the mean’. The emergence of drug-specific side effects can also ‘unblind’ a study. and they had higher response and remission rates than the rest (Wisniewski et al. 2003). there are measurement-related errors and biases.4 STUDIES CONSIDERED FOR REVIEW – ADDITIONAL INCLUSION CRITERIA In addition to the criteria established for the inclusion of trials for the guideline update as a whole. Despite the limitations of RCTs described above. the following specific criteria relating to RCTs of pharmacological treatments were established by the pharmacology topic group. 9. 2003). 2002). The inclusion of individuals likely to improve. participants may not be representative of patients seen in clinical practice because they are recruited by newspaper advertisement and paid for their participation in the study after completing a screening questionnaire (Greist et al. whatever they are given.). in which all participants are included using their last recorded value or assuming they have not improved. they are not blind to the phase of study. will all reduce the size of the specific drug effect. introduces potential bias the other way. The pressure to recruit patients may lead to ‘rating scale inflation’.Introduction to pharmacological and physical interventions Finally. In addition. A further consideration is that the method of analysis and confining the study to ‘completers’ may increase apparent drug effects. 2001. as well as those motivated to receive free medication. In addition to the factors related to the type of patient recruited into RCTs. and the fact that the same patients may appear in several publications (op. Melander et al. the tendency of journal editors to publish only studies with positive results (Kirsch & Scoboria. there are few alternatives to using these data because better ways of assessing efficacy have not been developed. 2002.

expert opinion and involvement with user groups. Studies were included provided data from the HRSD and MADRS could be extracted for the following outcomes: ● the number of participants who remitted75 (achieved below the equivalent 17-item HRSD score of eight) ● the number of participants who responded76 (achieved at least a 50% reduction in scores) ● mean endpoint or change scores in the rating scales. 1999. The GDG considered that these trials would still have adequate validity for determining efficacy in depression. studies were included provided there was clear evidence that at least 75% of patients received the standard dose or the mean dose used was at least 105% of the standard dose. relative risks for this outcome are framed in terms of the number of participants not responding. Furukawa et al.4.. Trials where participants had a diagnosis of atypical depression or depression with a seasonal pattern/seasonal affective disorder (SAD) were included provided all had a primary diagnosis of major depressive disorder. 76For statistical reasons.Introduction to pharmacological and physical interventions 9. 1971.1 Diagnosis Trials where some participants had a primary diagnosis of bipolar disorder were included provided at least 80% had a primary diagnosis of major depressive disorder and no more than 15% had a primary diagnosis of bipolar disorder. 1999).. Bollini et al. These figures resulted from discussion. Trials where some participants had a primary diagnosis of dysthymia were included provided at least 80% of trial participants had a primary diagnosis of major depressive disorder. In some situations where trial data were limited a greater proportion of patients with bipolar disorder were permitted but in this case the grade of evidence was reduced and these studies are identified. 2002a). 310 . 1989. Nevertheless. relative risks for this outcome are framed in terms of the number of participants not remitting..4. in order to provide fair comparisons. Trials not meeting these criteria are considered in the chapter on subthreshold depressive symptoms (Chapter 13). 75For statistical reasons. Thompson & Thompson.2 Dose There is a lack of clear evidence that doses of tricyclics at or below 100 mg are less effective than doses above (Blashki et al. although there might be benefit in more severely ill patients (Ramana et al. The standard dose was either that stated by Bollini and colleagues (1999) or by the BNF (2009) for drugs not included in Bollini and colleagues (1999).. and no more than 20% had a primary diagnosis of dysthymia. 9.

the pharmacology topic group had to decide which aspects of drug treatment were most important to clinicians and patients.2 Setting Where appropriate.5. Unfortunately there is little data to help with this point. acceptability and tolerability of the antidepressants most commonly prescribed in the UK. studies were categorised by setting: ● primary care – where this was specifically stated in the study ● outpatients/secondary care – where this was specifically stated in the study ● inpatients – where at least 75% of the patients were initially treated as inpatients. 9. this can be taken only as a guide to baseline severity because of heterogeneous samples with wide standard deviations as well as the fact that results are not presented in a way that allows differential response to be identified.5 ISSUES AND TOPICS COVERED BY THIS REVIEW In view of the vast numbers of studies performed investigating pharmacological responses in depression and the limited time available. Therefore the chapters on pharmacological interventions do not constitute a comprehensive review of all psychopharmacological studies performed in all aspects of the treatment of depression. the nature of the patients in these different groups varies.5. the cardiac safety of antidepressants.1 Severity A key issue is whether severity of illness can guide the use of antidepressant medication. Although the number of trial participants leaving treatment early was used as a measure of the tolerability of drugs 311 . In addition. although it is not clear how well ‘setting’ maps onto severity. This is likely to provide some bearing on the issue of setting and type of depression. dosage.3 Issues addressed In broad terms. 9. with depression with atypical features and with depression with psychotic symptoms.Introduction to pharmacological and physical interventions 9. Although most studies report mean baseline HRSD or MADRS. the GDG looked at some of the issues regarding so-called continuation and maintenance therapy. and issues regarding suicidality and completed suicide with antidepressants. A further problem is that because of differences among healthcare systems across the world. setting and age. Within each review. together with specific pharmacological strategies for dealing with depression that has inadequately responded to treatment.5. 9. Thus considerable uncertainty must be associated with conclusions drawn using these categories. the GDG looked at the effect on outcomes of severity. the GDG tried to address the issue of the comparative efficacy. where the data allowed.

pharmokinetics and pharmaceutical advice regarding these drugs. The following topics are covered: In the rest of this chapter: ● ● SSRIs versus placebo** (Sections 9.4 Topics covered Where there was lack of substantial new evidence.1): TCAs: amitriptyline** and overview of TCA data** (Section 10. some analyses and conclusions were not updated from the previous guideline (NCCMH. mirtazapine**. Readers are referred to conventional texts.8) ● St John’s wort** (Section 10. reboxetine** and venlafaxine** (Section 10.7) TCAs versus placebo (Section 9.4) The pharmacological management of atypical depression (Section 11. 312 . this guideline cannot be seen as a comprehensive review of the issue of the safety.3) The pharmacological management of depression with psychotic symptoms** (Section 11. 2004).3) ● Selective serotonin reuptake inhibitors (except escitalopram): citalopram**.6) Dosage issues for tricyclic antidepressants** (Section 11.10).5. Agomelatine was not licensed at the time of data analysis and is not included in this guideline. These are indicated with asterisks (**). Chapter 10: Pharmacological interventions Use of individual drugs in the treatment of depression (Section 10.7) ● ‘Third-generation’ drugs: duloxetine.8) The cardiotoxicity of antidepressants (Section 11.7) Antidepressant discontinuation symptoms** partly updated (Section 11.8). fluoxetine**. 9.5)** The physical and pharmacological management of depression with a seasonal pattern (Section 11. acceptability and tolerability for individual patients and drug interactions.Introduction to pharmacological and physical interventions reviewed.4) ● Escitalopram (Section 10. suicide and antidepressants** (Section 11.9). fluvoxamine**.9) Depression.5) ● Monoamine oxidase inhibitors: moclobemide**. although their discussion was updated where factual or stylistic adjustments were required. paroxetine** and sertraline** (Section 10. pharmacology.2) The effect of sex on antidepressant choice (Section 11. phenelzine** (Section 10.6 and 9. ● Chapter 11: Factors influencing choice of antidepressants ● ● ● ● ● ● ● ● ● The pharmacological management of depression in older adults** (Section 11. particularly those regarding issues of dosage schedules.

3.1) Switching to another antidepressant (Section 12. (2) measurement errors and biases. 2008)..3..3.9) – buspirone** (Section 12.10) – atomoxetine (Section 12. the response to placebo in an RCT consists of three main components: (1) spontaneous improvement. This means that the chance of improvement in response to antidepressants in clinical practice may not be the same as those in clinical trials involving placebo.4) – lithium (Section 12.3) ● Augmentation an antidepressant with a different drug.3.5) ● Relapse prevention** partly updated (Section 12. and (3) the true ‘placebo response’. so that the extent of the placebo response has been shown to correlate with the year of publication in studies in depression (r 0.3.6 PLACEBO-CONTROLLED RANDOMISED CONTROLLED TRIALS OF ANTIDEPRESSANTS As mentioned above. A recent meta-analysis showed that studies in which patients know they may get a placebo tablet have lower response rates than when they know they will only get active treatments (Sneed et al. evidence for the pharmacological treatment of symptoms of depression that do not meet threshold for major depressive disorder is considered in Chapter 13. which is non-pharmacological benefit due to taking part in a trial.4) ● TMS and VNS (Section 12.3.6). 2002). A large part of the placebo response is thought to be due to expectation combined with regular review and monitoring. Another systematic review provides suggestive evidence that the chance of responding to treatment with placebo is higher if monitoring is carried out more frequently in the first few weeks of treatment (Posternak & Zimmerman. including: – antipsychotics (Section 12.6) – pindolol** (Section 12. 2007).3.43) (Walsh et al.2) ● Combining an antidepressant with another antidepressant (Section 12. carbamazepine or valproate) (Section 12.5) – anticonvulsants** (lamotrigine. 9.7) – triiodothyronine (T3)** (Section 12.Introduction to pharmacological and physical interventions Chapter 12: The pharmacological and physical management of depression that has not adequately responded to treatment. Taking these factors together it is clear that the exact design of any trial will influence the non-specific benefit that participants will obtain and that the placebo response is not a minor distraction but an integral part of treatment not only in RCTs but also in clinical practice. In recent years there has been an increasing response to placebo.3.3. and relapse prevention: ● ● Increasing the dose (Section 12.3.11) ● ECT (Section 12.8) – benzodiazepines** (Section 12. In addition. There is a similar but less robust association between the extent of the response to active medication and the 313 .3.

One way round this problem is to use an active placebo. in Kirsch et al. However. and the drug–placebo difference becomes greater with increasing degrees of severity of depression (Angst. Khan et al.. An important point is that there is some evidence that the placebo response is greatest with mild depression.. Dago and Quitkin (1995) suggest that greater placebo response is more likely when the presenting episode occurs within the context of a psychosocial stressor. with more patients on placebo relapsing compared with those on antidepressants (Ross et al. Thus. 2002). less chronic disorders that have a greater chance of spontaneously improving or having a placebo response. 1986.. A meta-analysis of trials using active placebo is more effective than a metaanalysic of trials using only inactive placebos. 2008). probably resulting from subtly different procedures being adopted by different researchers (Schneider & Small. The increasing rate of response to placebo and to a lesser extent to antidepressants (Walsh et al... 2002) appears due to the decreasing efficacy of placebo with increasing severity of depression. 2008).). so that any effect of severity is diluted in group analyses. The published data did not allow the GDG to address the question of efficacy related to severity systematically since most RCTs 314 .. 1998. Kirsch et al.Introduction to pharmacological and physical interventions year of publication (r 0. although it is not clear how many of the trials involved have subsequently been published. the later work analysed all data submitted to the US Food and Drug Administration (FDA) for the licensing of new antidepressants.. an active drug effect becomes harder to detect (Thase. Kirsch et al. 2002). 2002) means that many trials are underpowered because with placebo response rates above 40%.26) (op. 2001). 2008). This effect cannot be demonstrated in the meta-analyses carried out for the guideline update because the published studies do not quote data for individual patients. there is considerable overlap between the distributions of HRSD scores between studies with different mean severities of depression at baseline. rather than increasing efficacy of the antidepressant drug per se (Kirsch et al. In three meta-analyses (Kirsch & Sapirstein. 2002a. The placebo response may also be short-lived. There is also the problem of ‘breaking the blind’ as a result of the side effects of antidepressants (Rabkin et al. that 80% of the antidepressant effect is placebo response.. 2002b) leading to possible bias in placebo-controlled clinical trials. including the SSRIs and venlafaxine. 2002.. there are few trials of active placebo using modern diagnostic criteria and widely accepted ratings (Moncrieff et al. 2002). Although the earlier meta-analysis was criticised because it included only a limited number of published trials. 1993.. 1993. cit.. Many commentators attribute this finding to placebo effects as discussed above. Angst. Kirsch et al. Khan et al.. This may well indicate an increasing tendency for RCTs to be carried out on people with milder. it has been argued that up to 80% of the antidepressant effect may be duplicated by placebo–that is. and between inpatient and outpatient studies. only for the entire group. Other methodological problems are highlighted by inter-site differences found in many multi-site trials. The increase in the drug/placebo difference with severity (Elkin et al. Longer trials are required to be able to fully elucidate the contributions of placebo and the treatment to clinical response. 1989.

each study considered for review is referred to by a ‘study ID’ made up of first author and publication date (unless a study is in press or only submitted for publication. is considered separately in Section 10. Improving these factors may help patients to cope with their daily lives thereby contributing to a reduction in symptoms of depression.1 Introduction The analysis of SSRIs as a class against placebo was not updated for this guideline although evidence for the most recently marketed SSRI. See Appendix 11 for a discussion of the severity categories used in the analyses in the previous guideline (in brief.7 SELECTIVE SEROTONIN REUPTAKE INHIBITORS VERSUS PLACEBO The following sections on SSRIs versus placebo marked by asterisks (**_**) are from the previous guideline and have not been updated except for style and minor clarification. Nonetheless. 78Here and elsewhere in the guideline. Study IDs in title case refer to studies included in the previous guideline and study IDs in capital letters refer to studies found and included in this guideline update. 315 .78 **One hundred and three studies were found in a search of electronic databases with 4879 being included and 55 being excluded by the GDG. 9. age and gender with relatively weak information about patient characteristics.5.Introduction to pharmacological and physical interventions merely give mean depression scores (with standard deviations) of large groups of patients. 9. when first author only is used). See the SSRIs versus placebo study charactertics table in Appendix 17c. 9. so that there is very considerable overlap between baseline depression scores of patients in different studies. 79This figure includes a multicentre trial (Kasper95) as well as two of its constituent trials published independently (Dominguez1985. the categories shift down so that moderate becomes mild. References for studies from the previous guideline are in Appendix 18 and references for studies for the update are in Appendix 17c.7. 77Details of standard search strings used in all searches are in Appendix 8.7. Information about each study along with an assessment of methodological quality is in Appendix 17c. the GDG’s findings were generally in favour of greater drug/placebo differences with increasing severity (see Section 9.2 Studies considered77. it was only possible to address important questions relating to the effects of severity. Lapierre1987) because ‘number of participants leaving the study early for any reason’ was not extractable from Kasper95. Therefore. It should also be borne in mind that there are non-mood-related benefits of prescribing antidepressants. for example in helping patients to sleep better and in dealing with anxiety-related symptoms. which also contains a list of excluded studies with reasons for exclusions. severe becomes moderate and very severe becomes severe).7). escitalopram.

Fabre1996. Edwards93. Fabre95. Feighner99. Lapierre1987. Stark85. Cohn1985. Sramek95.460 trial participants. Lapierre1987. 0. Conti1988. six as severe and nine as very severe. Walczak1996). Conti1988.69 to 0. McGrath00. Croft1999.65 to 0. a sub-analysis of studies which lasted 8 weeks or longer was undertaken. Feighner92. Mont’mery01. eight of paroxetine (Claghorn92a. Rickels1989. Three studies were of inpatients. Sil’stne99. with 16 trials of 8 weeks or longer. Wernicke1988). N 3143. 31 of outpatients.75 weeks). O’Flynn1991. it was not possible to determine which of the FDA data had been subsequently published.7. It was planned to combine these data with the FDA data reported by Kirsch and colleagues (2002a). 81Fifteen 316 . Feighner89a. RR 0. 95% CI. Norton1984. Dominguez85. In moderate82 depression there is some evidence suggesting that there is a clinically important difference favouring SSRIs over placebo on increasing the likelihood of patients achieving a 50% reduction in symptoms of depression as measured by the HRSD (K 383. 83Studies were excluded from sub-analyses of severity if mean baseline scores were not available. Hackett96. Coleman01. Rickels1992. studies were excluded from all efficacy outcomes because 50% left treatment early (Claghorn1996. Mendels1999.75. Rickels1986. Stahl00. Fabre1996. Feighner1989. Roth90. Walzak1996). Since it is possible that a placebo response is only short-lived. and were between 4 and 24 weeks’ long (mean 6. Visual inspection of funnel plots of the meta-analyses of the above studies indicated the possibility of publication bias. with four being classified as moderate. Dunlop1990. Lydiard1989. These provided data from up to 7. Dominguez85.87). 9. one in primary care and 13 either mixed or unspecified. However. RR 0. Mont’mery92A. N 729. Smith1992. Thakore1995. Stahl00). Cohn1985. Ravindram1995. Stark85. 95% CI. In severe depression there is strong evidence suggesting that there is a clinically important difference favouring SSRIs over placebo on increasing the likelihood of 80The forest plots can be found in Appendix 19c. Byerley88. In no study were more than 80% of study participants aged 65 years or over. Feighner92. Valducci1992. Feighner1989.78). Edwards93.Introduction to pharmacological and physical interventions Six studies were of citalopram (Burke02. 17 of fluoxetine (Andreoli2002. 12 of fluvoxamine (Claghorn1996. Itil1983.73. Rudolph99. Wernicke1987. Smith1992) and five of sertraline (Coleman1999. 82Severity categories based on APA (2000a) – see previous guideline Appendix 13. Reimherr90). Miller1989. Itil1983.3 Clinical evidence statements80 Effect of treatment on efficacy outcomes There is strong evidence suggesting that there is a clinically important difference favouring SSRIs over placebo on increasing the likelihood of patients achieving a 50% reduction in symptoms of depression as measured by the HRSD (K 1781. 0. Fabre95. Kasper95. It was possible to determine baseline severity in 19 studies. All included studies were published between 1983 and 2003.

0.83 to 0. RR 1. 0.28. N 866. There is insufficient evidence to determine whether there is a clinically important difference between SSRIs over placebo on increasing the likelihood of achieving remission as measured by the HRSD (K 3. 95% CI. Random effects RR 0. RR 0. 95% CI. There is some evidence suggesting that there is a clinically important difference favouring placebo over SSRIs on reducing the likelihood of patients reporting side effects (K 11. RR 0.8. 0. 95% CI.73).08 to 2.45. SMD 0.8).48 to 0.61. trials lasting 8 weeks or longer were analysed separately.94.47 to –0. 2. 95% CI. In very severe depression there is evidence suggesting that there is a statistically significant difference favouring SSRIs over placebo on reducing symptoms of depression as measured by the HRSD. 95% CI.13 to 1.54 to 0. 95% CI. 0.99).4). 95% CI.34. N 386. In moderate depression there is evidence suggesting that there is a statistically important difference favouring SSRIs over placebo on reducing symptoms of depression as measured by the HRSD. 0.63.89). 95% CI.54 to 0.24).08). 1. There is evidence suggesting that there is a statistically significant difference favouring SSRIs over placebo on reducing symptoms of depression as measured by the HRSD. 95% CI. 0.61 to 1.19.25). 84One study (Cohn1985) was removed from the meta-analysis to remove heterogeneity from the dataset. In severe depression there is some evidence suggesting that there is a clinically important difference favouring SSRIs over placebo on reducing symptoms of depression as measured by the HRSD (K 4. N 619.72. N 468. SMD 0. Random effects SMD 0.65 to 0. RR 0. but the size of this difference is unlikely to be of clinical importance (K 3984.22). SMD 0. In very severe depression there is strong evidence suggesting that there is a clinically important difference favouring SSRIs over placebo on increasing the likelihood of patients achieving a 50% reduction in symptoms of depression as measured by the HRSD (K 6.Introduction to pharmacological and physical interventions patients achieving a 50% reduction in symptoms of depression as measured by the HRSD (K 5. N 2290.88 to 0. RR 2. but the size of this difference is unlikely to be of clinical importance (K 2. There is strong evidence suggesting that there is a clinically important difference favouring placebo over SSRIs on reducing the likelihood of leaving treatment early due to side effects (K 39.06). N 7274. N 7460. 0. 317 . 95% CI. N 344. N 726. but the size of this difference is unlikely to be of clinical importance (K 16. 0. N 2223.39. but the size of this difference is unlikely to be of clinical importance (K 5. Sub-analysis of trials lasting 8 weeks or longer To assess whether the placebo effect was short-lived. Acceptability and tolerability of treatment There is evidence suggesting that there is a statistically significant difference favouring placebo over SSRIs on reducing the likelihood of leaving treatment early.

N 535. but the size of this difference is unlikely to be of clinical importance (K 1. In trials lasting 8 weeks or longer. 0. 0. there is evidence suggesting that there is a statistically significant difference favouring SSRIs over placebo on reducing symptoms of depression as measured by the HRSD. N 729.28. Random effects SMD 0. N 1369. RR 0.53.48 to 0. 0. there is strong evidence suggesting that there is a clinically important difference favouring SSRIs over placebo on increasing the likelihood of achieving a 50% reduction in symptoms of depression as measured by the HRSD (K 8. Acceptability and tolerability of treatment in trials lasting 8 weeks or longer In trials lasting 8 weeks or longer. In severe depression in trials lasting 8 weeks or longer. there is evidence suggesting that there is no clinically important difference between SSRIs and placebo on reducing the likelihood of 318 . N 386. 95% CI. In moderate depression in trials lasting 8 weeks or longer. 0.66 to 0.74). SMD 0. In trials lasting 8 weeks or longer. 0.59 to 0. RR 0. 95% CI.28. 0. 95% CI. N 283.67 to 1.Introduction to pharmacological and physical interventions Effect of treatment on efficacy outcomes in trials lasting 8 weeks or longer In trials lasting 8 weeks or longer. 95% CI.44 to 0. there is some evidence suggesting that there is a clinically important difference favouring SSRIs over placebo on increasing the likelihood of achieving a 50% reduction in symptoms of depression as measured by the HRSD (K 3. SMD 0. N 299. In severe depression in trials lasting 8 weeks or longer. there is evidence suggesting that there is a statistically significant difference favouring SSRIs over placebo on reducing symptoms of depression as measured by the HRSD. N 456.65 to 0.72. there is some evidence suggesting that there is a clinically important difference favouring SSRIs over placebo on reducing symptoms of depression as measured by the HRSD (K 1.43. 95% CI.2). 0.53 to 0. 0.79). RR 0. In very severe depression in trials lasting 8 weeks or longer. 95% CI. 95% CI.88). but the size of this difference is unlikely to be of clinical importance (K 2.87). RR 0. SMD 0. RR 0. N 1764.67 to 0.79 to 0. In very severe depression in trials lasting 8 weeks or longer.08). 95% CI. In moderate depression in trials lasting 8 weeks or longer.11). N 237.75. 95% CI. there is evidence suggesting that there is a statistically significant difference favouring SSRIs over placebo on reducing symptoms of depression as measured by the HRSD but the size of this difference is unlikely to be of clinical importance (K 7.07). 0. there is some evidence suggesting that there is a clinically important difference favouring SSRIs over placebo on increasing the likelihood of achieving a 50% reduction in symptoms of depression as measured by the HRSD (K 1.85.27).63. there is strong evidence suggesting that there is a clinically important difference favouring SSRIs over placebo on increasing the likelihood of achieving a 50% reduction in symptoms of depression as measured by the HRSD (K 3. there is insufficient evidence to determine whether there is a clinically important difference between SSRIs and placebo on increasing the likelihood of achieving remission as measured by the HRSD (K 2.72.

RR 1.09). 86The wording has been updated here. 0.4 Clinical summary There is strong evidence that SSRIs have greater efficacy than placebo on achieving a 50% reduction in depression scores in moderate and severe major depression85.069. There is some evidence for a similar effect in mild depression86.95.Introduction to pharmacological and physical interventions leaving treatment early (K 13. there is strong evidence suggesting that there is a clinically important difference favouring placebo over SSRIs on reducing the likelihood of leaving treatment early due to side effects (K 13.03).93.7.23 to 3. SSRIs produced more side effects than placebo.8. there is evidence suggesting that there is a statistically significant difference favouring placebo over SSRIs on reducing the likelihood of patients reporting side effects. There appears to be no difference between SSRIs and placebo on mean endpoint or change scores. but the size of this difference is unlikely to be of clinical importance (K 7.03 to 1. but for the guideline update these analyses were undertaken. N 1378. studies with statistically significant findings are more likely to be published than those with non-significant findings). Random effects RR 0. their utility in everyday clinical practice. There is insufficient evidence on the effect on remission because of heterogeneity in the meta-analysis. 319 . 95% CI. but the trend is towards a small effect size. Random effects RR 1. This was also the case in trials lasting 8 weeks or longer. 85The wording has been updated here. 1. 1. In trials lasting 8 weeks or longer. The previous guideline used the terms ‘severe and very severe depression’.1 TRICYCLIC ANTIDEPRESSANTS VERSUS PLACEBO Introduction In the previous guideline. a review of the efficacy and tolerability of TCAs compared with placebo was not carried out. The previous guideline used the term ‘moderate depression’. with more people leaving treatment early because of adverse events. therefore. This review informs the assessment of the relative efficacy and tolerability of different classes of antidepressants and.16). The effect was similar in longer trials.8 9. These results should be treated with caution because of publication bias (that is. 95% CI. N 3069. 95% CI.83 to 1. In trials lasting 8 weeks or longer. N 3.09. 9.** 9.

3 Studies considered87 In total. ICD or similar criteria Any TCA with UK marketing authority where a comparison with placebo was available 9.4 Clinical evidence Evidence from the important outcomes and overall quality of evidence are presented in (Table 66 and Table 67).Introduction to pharmacological and physical interventions Table 64: Databases searched and inclusion/exclusion criteria for clinical effectiveness of pharmacological treatments Electronic databases Date searched Update searches Study design Population Treatments MEDLINE. EMBASE. CINAHL Database inception to January 2008 July 2008. Information about the databases searched and the inclusion/exclusion criteria used are presented in Table 64. and little data for some outcomes from studies of clomipramine. 87Study IDs in capital letters refer to studies found and included in this guideline update. 9. respectively. 108 studies were found that met inclusion criteria. January 2009 RCT People with a diagnosis of depression according to DSM.2 Databases searched and the inclusion/exclusion criteria A systematic search for RCTs comparing any TCA with UK marketing authorisation with placebo was undertaken. The number of studies is summarised in Table 65.8. dosulepin and nortriptyline (see Table 66). PsycINFO. 9.8. The full evidence profiles and associated forest plots can be found in Appendix 16c and Appendix 19c. 320 . Most were for imipramine (66) and amitriptyline (30).8. Details of the search strings used are in Appendix 8. There were no extractable data from studies of lofepramine. with full details in Appendix 17c. which also includes details of excluded studies.

BYERLEY1988. FABRE1980. COHN1996. ITIL1993. STASSEN1993. CLAGHORN1996A. KLIESER1988. HOSCHL1989. HORMAZABAL1985. FONTAINE1994. ROWAN1982. COHN1992. FEIGHNER1979. FEIGHNER1992B. PECKNOLD1976B. SPRING1992. FEIGHNER1980. HOLLYMAN1988. SMITH1990. BEASLEY1991B. FEIGHNER1983A. BREMNER1995. FEIGHNER1989B. MYNORSWALLIS1997. HAYES1983. GOLDBERG1980. RICKELS1985. FABRE1996. FEIGER1996A. MYNORSWALLIS1995. GELENBERG1990. FEIGHNER1989. KASPER1995B. LECRUBIER1997B. LAPIERRE1991. ESCOBAR1980. KELLAMS1979. DOMINGUEZ1981. COHN1990A. RAMPELLO1991 FERGUSON1994B. BAKISH1992C. LYDIARD1989. FABRE1992. CLAGHORN1983. ROFFMAN1982. CLAGHORN1983B. DUNBAR1991. RICKELS1982D. THOMPSON2001B BARGESCHAAPVELD2002. LAAKMAN1995.Introduction to pharmacological and physical interventions Table 65: Summary of studies for TCAs versus placebo TCA Amitriptyline Number of studies 30 Study IDs AMSTERDAM2003A. LAIRD1993. REIMHERR1990. FEIGHNER1989C. Continued 321 Clomipramine Dosulepin Imipramine 3 4 66 . COHN1985. FEIGHNER1983B. RICKELS1991. ELKIN1989. GELENBERG1990. ENTSUAH1994. GEORGOTAS1982A. BAKISH1992B. COHN1984. LYDIARD1997. MINDHAM1991. CASSANO1996. HICKS1988. GERNER1980B. LAPIERRE1987. FEIGHNER1989A. BOYER1996A. DOMINGUEZ1985. WILCOX1994 LARSEN1989. ITIL1983A. FEIGHNER1993. FEIGHNER1982. CASSANO1986. LIPMAN1986.

WAKELIN1986 Lofepramine Nortriptyline Total 0 4 108 N/A GEORGOTAS1986. PESELOW1989. VERSIANI1990. SILVERSTONE1994. PEDERSEN2002. WHITE1984A On all measures of efficacy TCAs are more effective than placebo. However. There was little difference between TCAs compared with placebo for leaving treatment early. there was only a single study. apart from clomipramine which only showed a similar result for number of participants reporting side effects. However. RICKELS1987. NORTON1984. PESELOW1989B. SMALL1981. dosulepin and clomipramine). Results were similar for each individual drug where there were sufficient data. QUITKIN1989. MARKOWITZ1985. SCHWEIZER1994.Introduction to pharmacological and physical interventions Table 65: (Continued) TCA Number of studies Study IDs MARCH1990. RICKELS1981. RICKELS1982A. SCHWEIZER1998. UCHA1990. although effect sizes were less certain for individual drugs with few data (for example. participants taking TCAs were more likely to leave treatment early because of side effects and to report side effects than those taking placebo. 322 . NANDI1976. MERIDETH1983. SHRIVASTAVA1992. PHILIPP1999. NAIR1995. MENDELS1986. KATZ1990. This finding was similar across individual drugs.

61 ( 0. 386 K n 13. K participants n 22.41 ( 0.5 versus 67.4) Quality Moderate Number of studies.37) ( 0. 2445 Forest plot number Pharm TCAs 01.62 to 0.01 Mean depression change scores at endpoint High K n Pharm TCAs 01.Table 66: Summary evidence profile for TCAs versus placebo (efficacy data) Amitriptyline Not reported SMD 0.49 ( 0.68 Not reported (0.7 to 0.5 ( 0.62 to 0.5 versus 65.01 Pharm TCAs 01.18) ( 0.03 Not reported Overall (all studies) Mean depression scores at endpoint High K n Pharm TCAs 01.24) Moderate K n 2.48 SMD 0.65 to 0.34) Quality Moderate – – – Number of studies.71 (0.41 to 0.4%) Introduction to pharmacological and physical interventions Continued 323 .35 SMD 0. 645 – – – – SMD 0.6%) (41.01) High K n – 4.03 Not reported – 3.74 RR 0.67 to 0.66 to 0. 1603 Moderate SMD 0.88) (0.3%) RR 0.54 to 0.8 ( 1.70 RR 0.01 Not reported – 6.78) (43 versus 62.7%) (43. 528 Pharm TCAs 01.75) (0.53 to 0.03 Non-response RR 0. K participants n 7.01 Not reported 1.8 versus 63.37 to 0. 348 – – SMD 0.83 to 0.76) (48.21 ( 0.29) Moderate K n Pharm TCAs 01. 1173 Forest plot number Pharm TCAs 01.59 to 0.01 SMD 0.27) Clomipramine Dosulepin Imipramine Nortriptyline SMD 0. 108 Pharm TCAs 01.

131 Pharm TCAs 01.06 Pharm TCAs 01.48) (45 versus 77.8 versus 10%) Moderate K n Pharm TCAs 01.66 (0.8%) (11.91) (0. 4717 Forest plot number Pharm TCAs 01.58 RR 1.74 RR 0.88) (70.83 RR 0. 954 Forest plot number Pharm TCAs 01.4%) (62. 386 K n 20.18 (0.75 to 0.04 .18 to 7.34 to 1) (0.65 to 0. K participants n 09. 2145 – K n 1.68 (0. 152 Overall (all studies) Quality High Number of studies. 596 Pharm TCAs 01.06 Non-remission RR 0.84) (0.06 13.04 Moderate K n 2.04 Moderate – Pharm TCAs 01.3%) Moderate K n 2.04 1.1 versus 92.04 Moderate K n Pharm TCAs 01.04 3.52 to 0. K participants n 34.06 RR 0.1%) Introduction to pharmacological and physical interventions Quality Moderate Number of studies. 2186 – High Moderate – – – Clomipramine Dosulepin Imipramine Nortriptyline RR 0.4 versus 85. 38 K n 1.324 Table 66: (Continued) Amitriptyline High K n Pharm TCAs 01.9 versus 83. 37 Pharm TCAs 01.44 to 1) (63 versus 82.6%) (51.

9%) Moderate K n 19.02 16.93 (0.2 versus 1.03 to 1.92 (0.09) (18.2) (80 versus 50%) Low K n 1.6%) Quality High Number of studies.6 versus 34. 475 K n 53.44 (1. 4523 325 Forest plot number Pharm TCAs 02. 2350 K n 1.03 .99 (0. 251 Pharm TCAs 02.44) (16.98 (1.4 versus 46%) Clomipramine Dosulepin Imipramine Nortriptyline Summary evidence profile for TCAs versus placebo (leaving treatment early and side-effect data) Overall (all studies) Leaving treatment early RR 0.79 to 1. 129 Pharm TCAs 02.01 Leaving treatment early due to side effects RR 4.5%) High K n 2.73 (0.67) (18.6%) (75.74) (1.15 to 1.03 23.01 Pharm TCAs 02. 409 Low High RR 0. 38 K n 2.5%) Moderate K n 1.47 to 5.03 RR 1.27 to 4. 52 Pharm TCAs 02.1) (32.5) (40.3 to 2.1%) (14.14 to 5.66 (3.02 RR 3.09 (0.02 Number reporting side effects Moderate K n Pharm TCAs 02.2 versus 57.19) (20 versus 25%) RR 1. 6288 K n Moderate Moderate Moderate Low 3.62) (74.79) (74. 113 Pharm TCAs 02.5 versus 81%) High K n 2.38 to 6. 932 RR 1.01 RR 4.4 RR 1.1%) High K n 44.02 Pharm TCAs 02.7 versus 51%) Quality Moderate Introduction to pharmacological and physical interventions Number of studies.7 versus 3.03 RR 0.06) (37 versus 37.9 RR 2. 2805 K n 2.03) (35.18 (1.92 to 1.6 (0.03 Pharm TCAs 02.22 to 1. 8173 Forest plot number Pharm TCAs 02.01 RR 3. 3390 Pharm TCAs 02.7 versus 4.3%) Moderate K n Pharm TCAs 02.58) (1.8%) RR 0.9 versus 38.02 (1.01 Pharm TCAs 02.93 to 1.67) (20 versus 5.09) (38.34) (95. 58 K n 3.79 to 1.5%) High K n Pharm TCAs 02.82 (0.46 to 4.02 (3.03 7.01 RR 7.8 to 3. participants K n 30.18) (56 versus 18. 5245 Pharm TCAs 02.3%) RR 1.02 RR 1.Table 67: Amitriptyline RR 0.01 (0.8) (10 versus 11. 20 Pharm TCAs 02.26 to 1. participants K n 84. 9901 Forest plot number Pharm TCAs 02.41 (1. participants K n 65.7 versus 39.5 versus 5%) Pharm TCAs 02.51 to 42.6%) Quality Moderate Number of studies.27 to 2.9 versus 56.91 (3.27 to 7.02 RR 1.

When compared with the review of SSRIs against placebo. there were only five studies in the analysis. which is not enough to draw conclusions.2.40 –1.81 22.Introduction to pharmacological and physical interventions Figure 8: Meta-regression showing relationship between baseline depression scores and effect sizes calculated from mean endpoint or mean change scores Regression of Baseline depression score on Hedges's g 0.60 Hedges's g –0.20 –0. However. and it is not reported here. and similar with regard to completing treatment.47 31. However. This showed no consistent relationship between baseline scores and effect sizes calculated from mean endpoint depression score or change score (regression coefficient 0. but larger on mean endpoint data than those seen with SSRIs.6 Clinical summary TCAs are more effective than placebo in terms of efficacy. Section 10.00 –0.31 16. 9. Sensitivity analyses for mean endpoint scores and mean change scores were performed and a relationship found between mean change scores and baseline scores. they are more likely to lead to stopping treatment due to side effects and more likely to cause side effects.14 14.80 –1.8.48 18.20 –1.14 27.00 12. This may be explained by the fact that the included studies were mostly older than those in the SSRI review and the differences in effect sizes may be explained by a combination of the timing of the studies and the characteristics of the participants. the effect sizes from efficacy outcomes tended to be similar for response outcomes.64 33.80 –2.46]).8.97 25.31 29.00 –1.60 –1.4.5 Effect of baseline severity on outcomes A meta-regression was undertaken using the baseline depression scores as the predictor variable (see Figure 8). A review of SSRIs compared with TCAs is in Chapter 10. 326 .81 Baseline depression score 9.01 [p 0.40 –0.64 20.

1. The previous guideline recommended that antidepressants should not be prescribed for mild depression based on the poor risk–benefit ratio.9 FROM EVIDENCE TO RECOMMENDATIONS There is evidence that antidepressants are more effective than placebo on efficacy outcomes. There is some evidence that they are less effective in people with less severe symptoms.10. 9.10 9. but consider them for people with: ● a past history of moderate or severe depression or ● initial presentation of subthreshold depressive symptoms that have been present for a long period (typically at least 2 years) or ● subthreshold depressive symptoms or mild depression that persist(s) after other interventions88. 88This recommendation also appears in Chapter 13. with those taking TCAs more likely to report side effects or leave treatment early because of side effects. but could be considered for persistent symptoms following other interventions or for those with a history of moderate or severe depression.1 RECOMMENDATION Do not use antidepressants routinely to treat persistent subthreshold depressive symptoms or mild depression because the risk–benefit ratio is poor. 327 . this recommendation may be extended to include the TCAs. but may be effective in persistent subthreshold depressive symptoms (dysthymia). 9.Introduction to pharmacological and physical interventions The effect sizes for tolerability outcomes were considerably larger than those seen with SSRIs. Given the evidence in Chapter 13 reviewed for the guideline update. that antidepressants are not more effective than placebo for recent onset subthreshold depressive symptoms. but that they are less acceptable (based on attrition rates). and produce more side effects.

there are few medium-term data (Anghelescu et al. Nemeroff et al.. 2006. The GDG updated its reviews of drugs (including escitalopram and antipsychotic augmentation) where there were substantial new data likely to change the recommendations from the previous guideline. Although new RCT data on venlafaxine have become available and several meta-analyses (for example. 2008).Pharmacological interventions 10 10. these new data do not change the conclusion that if there is an efficacy advantage for venlafaxine over other antidepressants it is small and unlikely to be of clinical importance. The GDG did not update its review of St John’s wort. the GDG concluded that there was sufficient doubt about the clinical importance of the differences to not justify the development of recommendations for specific drugs.. and where studies for newly licensed drugs (duloxetine) were available. 2008) or data that support the use of St John’s wort in relapse prevention (Kasper et al. There is also a lack of efficacy data in people with severe depression and long-term safety data remain scant.. differences between drugs relating to tolerability and safety are highlighted where relevant.11. Some of the recommendations were revised (NICE. 2006b). Although further data have become available to suggest that St John’s wort may be more effective and better tolerated than standard antidepressants in the acute treatment of mild to moderate depression. 2008. 2008). This includes SSRIs (apart from escitalopram) and venlafaxine.. 2009). A recent network meta-analysis has also been published (Cipriani et al. However... 2007a) in light of the safety review of venlafaxine conducted by the Medicines and Healthcare products Regulatory Agency (MHRA 2006a. and further revised in this guideline update. 2008) and systematic reviews (Gartlehner et al. Kasper et al. Weinmann et al. Cipriani et al. The relative efficacy and tolerability of SSRIs and serotonin–noradrenaline reuptake inhibitors (SNRIs) have been the subject of several meta-analyses (for example. this is discussed in more detail in Section 10.. These analyses do suggest that there may be differences in efficacy and tolerability between individual drugs but. Gartlehner et al.. which uses direct and indirect methods to rank 12 new antidepressants with regard to relative efficacy and tolerability. In addition. 2008).. It did not update most of the reviews of individual antidepressants undertaken for the previous guideline because most of these were large-scale reviews – a substantial amount of new evidence would have had to have been published to change the overall conclusion that there is little difference in efficacy between individual drugs. The GDG 328 .. there is evidence of publication bias that complicates the interpretation of these data (Linde et al. 2008) have been published. given the modest size of the effect and some methodological uncertainties.1 PHARMACOLOGICAL INTERVENTIONS INTRODUCTION This chapter reviews the use of individual drugs in the treatment of depression. 2008.

TCAs included clomipramine. some analyses and conclusions were not updated from the previous guideline (NCCMH.2 USE OF INDIVIDUAL DRUGS IN THE TREATMENT OF DEPRESSION Where there was lack of substantial new evidence. 10.5) 89Many studies in the reviews used a TCA as a comparator treatment. The recommendations on St John’s wort remain.2.Pharmacological interventions were previously cautious about the use of St John’s wort partly because there is uncertainty over the active constituent and the majority of preparations are not standardised to contain fixed quantities of individual constituents. It covers the following drugs: ● Tricyclic antidepressants (TCAs) (Section 10. 2004). the effect of patient setting (inpatient. The reviews of escitalopram and duloxetine are new for this guideline update. nortriptyline. outpatient or primary care) on choice of drug was also examined. Traditional Herbal Registration Certificates have been granted in the UK for standardised preparations of St John’s wort. These data were combined in a review of TCAs to enable the GDG to gain an overview of this class of drugs. dothiepin/dosulepin.1 Introduction This section reviews the relative efficacy of individual antidepressants in the treatment of depression. 329 . these certificates are not based on RCT evidence of efficacy and tolerability in the same way that a product licence is for a conventional medicine. These are indicated with asterisks (**). 10. desipramine. amineptine and lofepramine.3) – Amitriptyline** – An overview of TCAs used as comparator treatments in trials reviewed elsewhere89** ● Selective serotonin reuptake inhibitors (SSRIs) except escitalopram (Section 10. imipramine. Since the previous guideline was published. although their discussion was updated where factual or stylistic adjustments were required. Where there were sufficient data.4) – Citalopram** – Fluoxetine** – Fluvoxamine** – Paroxetine** – Sertraline** ● Escitalopram (Section 10. doxepin. unchanged. therefore.

7) – Moclobemide** – Phenelzine** ‘Third-generation’ drugs (Section 10. with seizures and arrhythmias being a particular concern (see Chapter 11. In many patients a ‘therapeutic dose’ is never reached either because the patient cannot tolerate it or because the prescriber does not titrate the dose upwards.Pharmacological interventions ● ● ● Monoamine oxidase inhibitors (Section 10. sedation and postural hypotension.3 TRICYCLIC ANTIDEPRESSANTS The following sections on TCAs marked by asterisks (**_**) are from the previous guideline and have not been updated except for style and minor clarification.9). constipation. to varying degrees. This toxicity and the perceived poor tolerability of these drugs in general have led to a decline in their use in the UK over the last decade. blurred vision. clomipramine being primarily serotonergic and imipramine noradrenergic. and sweating). Efficacy benefits may be more marked in hospitalised patients (Anderson et al. Currently. 10. thus enhancing noradrenergic and serotonergic neurotransmission.2 Amitriptyline Although amitriptyline was not the first TCA and is not the best tolerated or the most widely prescribed. a potential benefit that is offset by its poorer tolerability (Barbui & Hotopf. it is the standard drug against which new antidepressants are compared with respect to both efficacy and tolerability. All TCAs except lofepramine are toxic in overdose.8) – Duloxetine – Mirtazapine** – Reboxetine** – Venlafaxine** Other preparations – St John’s wort** (Section 10. nine TCAs are available in the UK. Amitriptyline may be marginally more effective than other antidepressants. urinary retention.1 Introduction **TCAs have been used to treat depression for over 40 years.3. Section 11. 10.. Although all TCAs block the reuptake of both amines. anticholinergic side effects (dry mouth.9) 10. they vary in their selectivity with. They are thought to exert their therapeutic effect by inhibiting the re-uptake of monoamine neurotransmitters into the presynaptic neurone. 2001). 2000). 330 . for example.3. All TCAs cause. These side effects necessitate starting with a low dose and increasing slowly.

The original systematic review on which this section is based included two outcome measures. Preskorn1991. Sixteen studies were of inpatients. minaprine93. Information about each study along with an assessment of methodological quality is in Appendix 17c. amoxapine. In the remaining eight.71 weeks). when first author only is used). 2001) as the basis for this section. maprotiline. which also contains a list of excluded studies with reasons for exclusions. sertraline. All included studies were published between 1977 and 1999 and were between 3 and 10 weeks’ long (mean 5. responders and mean endpoint scores. 331 . it was either not clear from where participants were sourced or they were from mixed sources. Study IDs in title case refer to studies included in the previous guideline and study IDs in capital letters refer to studies found and included in this guideline update. Shaw1986. 95The authors of the review on which this review is based entered data into Review Manager so that amitriptyline is on the right-hand side of the forest plot and comparator treatments on the left. doxepin. Hutchinson92). fluoxetine. lofepramine. 93Not available in the UK. dothiepin/dosulepin. Stuppaeck1994. 22 of outpatients and two were undertaken in primary care. References for studies from the previous guideline are in Appendix 18 and references for studies for the update are in Appendix 17c. Guy1983. desipramine. Geretsegger95. each study considered for review is referred to by a ‘study ID’ made up of first author and publication date (unless a study is in press or only submitted for publication. paroxetine.48492 participants and efficacy data from up to 2. nortriptyline. trimipramine. In three studies all participants were over the age of 65 years (Cohn1990. The most common reason for exclusion was an inadequate diagnosis of depression. 91Here and elsewhere in the guideline. fluvoxamine. for which the authors made their data available to the NCCMH team. therefore. Fawcett1989. the figures given are for the outcome with the largest number of participants. The original review included 184 studies of which 144 did not meet the inclusion criteria set by the GDG. 96Where it made a difference to results the following studies were removed from efficacy analyses because 50% left treatment early: Cohn1990. Eight additional studies were identified from searches undertaken for other sections of this guideline. imipramine.95 Effect of treatment on efficacy96 There appears to be no clinically important difference in efficacy between amitriptyline and other antidepressants. either when compared or by class: There is evidence suggesting that there is no clinically important difference between other antidepressants and amitriptyline on increasing the likelihood of 90Details of standard search strings used in all searches are in Appendix 8. Clinical evidence statements for amitriptyline94. Data were available to compare amitriptyline with citalopram. Thus 48 trials are included in this section providing tolerability data from up to 4. Studies reported mean doses equivalent to at least 100 mg of amitriptyline. phenelzine and mirtazapine. trazodone. 92It is not always possible to extract data for all outcomes from each study. mianserin. Wilcox1994.760 participants. A total of 177 trials were excluded.91 The GDG used an existing review (Barbui & Hotopf. 94The forest plots can be found in Appendix 19c.Pharmacological interventions Studies considered90. It did not include data on remission and this has not been extracted for the present review.

95% CI. 95% CI.04. 95% CI. Effect of setting on treatment efficacy There appears to be no clinically important difference between amitriptyline and other antidepressants in different treatment settings: In inpatients there is evidence suggesting that there is no clinically important difference between other antidepressants and amitriptyline on increasing the likelihood of achieving a 50% reduction in symptoms of depression as measured by the HRSD (K 6. N 600. 0. RR 0. 0. There is insufficient evidence to determine whether there is a clinically important difference between other TCAs and amitriptyline on increasing the likelihood of achieving a 50% reduction in symptoms of depression as measured by the HRSD (K 2. but the size of this difference is unlikely to be of clinical importance (K 9. 95% CI.60 to 1.2).96. 0. –0. 95% CI. There is evidence suggesting that there is a statistically significant difference favouring amitriptyline over other antidepressants on reducing symptoms of depression by the end of treatment as measured by the HRSD and MADRS.89 to 1. N 1541. There is evidence suggesting that there is no clinically important difference between: ● other TCAs and amitriptyline on reducing symptoms of depression by the end of treatment as measured by the HRSD or MADRS (K 5.25) ● SSRIs and amitriptyline on reducing symptoms of depression by the end of treatment as measured by the HRSD or MADRS (K 19.19 to 0.01 to 0.03 to 0. RR 1.00 to 0.06. SMD 0. N 1002.06. N 285.53).29).27) ● SSRIs and amitriptyline on increasing the likelihood of achieving a 50% reduction in symptoms of depression as measured by the HRSD (K 9. 0. In outpatients there is evidence suggesting that there is no clinically important difference between other antidepressants and amitriptyline on increasing the likelihood of achieving a 50% reduction in symptoms of depression as measured by the HRSD (K 7. In outpatients there is evidence suggesting that there is a statistically significant difference favouring amitriptyline over other antidepressants on reducing symptoms of depression as measured by the HRSD and MADRS. RR 1. –0. SMD 0. SMD 0. 95% CI.08. SMD 0. SMD 0. 95% CI. 0.16). 0. but the size of this difference is unlikely to be of clinical importance (K 11. In inpatients there is evidence suggesting that there is a statistically significant difference favouring amitriptyline over other antidepressants on reducing symptoms of depression as measured by the HRSD and MADRS.03. 95% CI. N 68.30).02 to 0.18). N 2760. 95% CI. 95% CI. RR 1.09.09. RR 1. N 752. 0. but the size of this difference is unlikely to be of clinical importance (K 32.96 to 1.13.95 to 1. 0. N 666.25). N 837.Pharmacological interventions achieving a 50% reduction in depression scores as measured by the HRSD (K 16. N 1648.9 to 1.16.16). In patients in primary care there is evidence suggesting that there is no clinically important difference between other antidepressants and amitriptyline on reducing 332 .

27). 2.78. 0. N 855.44 to 0.78. There is some evidence suggesting that there is a clinically important difference favouring other antidepressants over amitriptyline on reducing the likelihood of outpatients leaving treatment early due to side effects (K 18.84 to 1. There is insufficient evidence to determine whether there is a clinically important difference between other antidepressants and amitriptyline on reducing the likelihood of inpatients leaving treatment early due to side effects (K 8.87.82 to 1.Pharmacological interventions symptoms of depression by the end of treatment as measured by the HRSD (K N 132. However.04). RR 0. There is evidence suggesting that there is no clinically important difference between paroxetine and amitriptyline on reducing the likelihood of inpatients reporting side effects (K 2. 0. There is some evidence suggesting that there is a clinically important difference favouring other antidepressants over amitriptyline on reducing the likelihood of patients reporting side effects (K 5. RR 0. N 2647.09. N 4884.8.72 to 1. there appears to be little difference between the tolerability of amitriptyline and other antidepressants: There is evidence suggesting that there is no clinically important difference between other antidepressants and amitriptyline on reducing the likelihood of inpatients leaving the study early for any reason (K 15.88.96. amitriptyline appears to be equally tolerable in terms of leaving treatment early for any reason. 95% CI. N 1320. RR 0. Acceptability and tolerability of treatment When compared with all antidepressants. RR 0.83). RR 0.61 to 0. patients taking other antidepressants report fewer side effects: There is evidence suggesting that there is no clinically important difference between amitriptyline and other antidepressants on reducing the likelihood of leaving treatment early for any reason (K 43. N 773. Acceptability and tolerability of treatment by setting For inpatients.13). N 131.06). 95% CI. 95% CI. SMD –0.9). There is insufficient evidence to determine whether there is a clinically important difference between other antidepressants and amitriptyline on reducing the likelihood of outpatients reporting side effects (K 2.75. RR 0. Random effects RR 0.1).92. 0. 0. 0. RR 0. There is strong evidence suggesting that there is a clinically important difference favouring other antidepressants over amitriptyline on reducing the likelihood of leaving the study early due to side effects (K 34.93). 0. RR 0.003). 95% CI. 95% CI. 95% CI. 95% CI. 95% CI.12). 0.61 to 1. N 2396. There is evidence suggesting that there is no clinically important difference between other antidepressants and amitriptyline on reducing the likelihood of outpatients leaving treatment early for any reason (K 19. 95% CI. N 552.62 to 0.65 to 0. 0. 0. N 4034. 333 . 95% CI. –0.55 to 1.68 to 1.71. Amitriptyline was less well tolerated in outpatients.

967 patients.35 to 0. All included studies were published between 1981 and 2002.55. amineptine and lofepramine.3 Tricyclic antidepressants – an overview of selected data **This section combines data from other reviews where a TCA was used as a comparator treatment. Information about each study along with an assessment of methodological quality is in Appendix 17c. It is.8. eight from phenelzine (Section 10. and tolerability data from up to 8. 48 of outpatients and three undertaken in primary care. which also contains a list of excluded studies with reasons for exclusions. N 90. dothiepin/dosulepin. 98Study IDs in title case refer to studies included in the previous guideline. participants had depression with additional atypical features. and. There is some evidence suggesting that there is a clinically important difference favouring paroxetine over amitriptyline on reducing the likelihood of primary care patients reporting side effects (K 1. It specifically does not include comparisons of TCAs with other TCAs. In 11 more than 80% of study participants were aged 65 years and over. 94 studies from other reviews included a TCA as a comparator drug. not a systematic review since a systematic search for all trials of TCAs was not conducted. seven from the review of mirtazapine (Section 10. more patients reported side effects in amitriptyline than paroxetine.4).3). therefore. Efficacy data were available from up to 6. which was the only comparator drug available: In patients in primary care there is insufficient evidence to determine whether there is a clinically important difference between other antidepressants and amitriptyline on reducing the likelihood of leaving treatment early either for any reason or due to side effects.5). three from reboxetine (Section 10. Data were available from the following TCAs: clomipramine.3). imipramine. 95% CI. RR 0. Seventy studies were sourced from the review of SSRIs (Section 10.8. 10. In the remaining 19. desipramine. 97Details of standard search strings used in all searches are in Appendix 8.3. 334 . References for these studies guideline are in Appendix 18. nortriptyline.8. although patients taking the drug report more adverse events and tend to leave treatment early due to side effects. doxepin. Clinical summary Amitriptyline is as effective as other antidepressants.848 patients.86). 0.7.4) and six from venlafaxine (Section 10. Twenty-four studies were of inpatients.98 In all. Studies considered97. it was either not clear from where participants were sourced or they were from mixed sources. in two.Pharmacological interventions Although much of the evidence was too weak to make a valid comparison of tolerability in primary care.

103Quitkin1990 was removed from the meta-analysis to remove heterogeneity from the imipramine dataset. 102Bruijn1996 was removed from the meta-analysis to remove heterogeneity from the imipramine dataset.64 to 0.01) ● increasing the likelihood of achieving remission as measured by the HRSD (K 3101. 0. N 1681. N 3. N 60. in inpatients by the end of treatment. 95% CI. as measured by the HRSD or the MADRS.91. 99The forest plots can be found in Appendix 19c. 95% CI. 95% CI.02. N 765. –0.22). N 733.275.03.82 to 1.12. There is insufficient evidence to determine whether there is a clinically important difference between phenelzine and nortriptyline on increasing the likelihood of achieving remission in outpatients by the end of treatment as measured by the HRSD (K 1103.09). RR 0. and Quitkin1990 were removed from the meta-analysis to remove heterogeneity from the imipramine dataset. 0. 95% CI. 95% CI. N 2364. –0.98. There is evidence suggesting that there is no clinically important difference between TCAs and alternative antidepressants on reducing symptoms of depression in outpatients by the end of treatment as measured by the HRSD or MADRS (K 33. 95% CI.74. RR 0. 0. 100Bruijn1996 335 . SMD 0. N 534.04).03 to 0. RR 0.848. SMD 0.98.03 to 0.15) ● reducing symptoms of depression by the end of treatment as measured by the HRSD or MADRS (K 70. N 6.84 to 1.07). 95% CI. Effect of setting on treatment efficacy Inpatients: There is evidence suggesting that there is no clinically important difference between TCAs and alternative antidepressants on increasing the likelihood of achieving a 50% reduction in symptoms of depression in inpatients as measured by the HRSD (K 4102. 0. 101Quitkin1990 was removed from the meta-analysis to remove heterogeneity from the imipramine dataset.Pharmacological interventions Clinical evidence statements99 Effect of treatment on efficacy There is evidence suggesting that there is no clinically important difference between other antidepressants and TCAs on: ● increasing the likelihood of achieving a 50% reduction in symptoms as measured by the HRSD or the MADRS (K 15100.87). RR 1. but the size of this difference is unlikely to be of clinical importance (K 20. 0.28.1 to 0. There is evidence suggesting that there is a statistically significant difference favouring TCAs over alternative antidepressants on reducing symptoms of depression.78 to 2.18). RR 0. 0. SMD –0.83 to 1. 95% CI. Outpatients: There is some evidence suggesting that there is a clinically important difference favouring alternative antidepressants over TCAs on increasing the likelihood of achieving a 50% reduction in symptoms of depression as measured by the HRSD (K 5.

42. 95% CI.** 10.96).02. 1.65 to 0.1 Introduction **The selective serotonin reuptake inhibitors (SSRIs) inhibit the reuptake of serotonin into the presynaptic neurone thus increasing neurotransmission. RR 0. they are not serotonin specific. When TCAs were examined individually. N 8888.02 to 1. 95% CI. There is strong evidence suggesting that there is a clinically important difference favouring alternative antidepressants over TCAs on reducing the likelihood of leaving treatment early due to side effects (K 80.94) ● on reducing the likelihood of patients reporting adverse effects (K 25. Although they ‘selectively’ inhibit serotonin reuptake.78). 95% CI. 0. N 3007.88. but the size of these differences is unlikely to be of clinical importance: ● on reducing the likelihood of leaving treatment early for any reason (K 83.71. RR 0. N 336. RR 2. only dothiepin/dosulepin appears to be more acceptable than alternative antidepressants: There is some evidence suggesting that there is a clinically important difference favouring dothiepin/dosulepin over alternative antidepressants on reducing the likelihood of leaving treatment early for any reason (K 5.13). Some of the drugs in this class also inhibit the reuptake of noradrenaline and/or dopamine to a lesser extent. random effects RR 0. As a class. 0. N 213.14.83 to 0.4. Clinical summary TCAs have equal efficacy compared with alternative antidepressants but are less well tolerated particularly in outpatients. Dosage titration is not routinely 336 .76). they are associated with less anticholinergic side effects and are less likely to cause postural hypotension or sedation.09 to 3. –0. Acceptability and tolerability of treatment There is evidence suggesting that there are statistically significant differences favouring alternative antidepressants over TCAs on the following outcomes.Pharmacological interventions Primary care: There is insufficient evidence to determine whether there is a clinically important difference between TCAs and alternative antidepressants on reducing symptoms of depression in patients in primary care by the end of treatment as measured by the HRSD or MADRS (K 2. 95% CI.98) and on reducing the likelihood of leaving treatment early due to side effects (K 5.91. RR 1. 95% CI. SMD –0. N 8967.86 to 0. N 336.4 SELECTIVE SEROTONIN REUPTAKE INHIBITORS The following sections on SSRIs marked by asterisks (**_**) are from the previous guideline and have not been updated except for style and minor clarification. 95% CI.42 to 0. 0. 10. 1.

It is associated with a higher incidence of nausea than the other SSRIs and so is not widely prescribed. while both citalopram and escitalopram are relatively safe in this regard. fluoxetine and paroxetine are potent inhibitors of various hepatic cytochrome metabolising enzymes (Mitchell. The most problematic side effects of this class of drugs are nausea.. They are also less cardiotoxic and much safer in overdose than the TCAs or MAOIs.3). There are other important differences among the SSRIs (Anderson & Edwards.. Escitalopram104 Citalopram is a racemic mixture of s-citalopram and r-citalopram. escitalopram (s-citalopram) is 100 times more potent than r-citalopram. which may cause problems with washout periods when switching to other antidepressant drugs but has the advantage of causing fewer discontinuation symptoms. one of its minor metabolites is cardiotoxic (Van Der Burght.Pharmacological interventions required so subtherapeutic doses are less likely to be prescribed. efficacy will be maintained and side effects reduced. diarrhoea and headache. In animals. Fluvoxamine. It has been suggested that r-citalopram contributes to side effects and.5. These advantages have led to their widespread use as better-tolerated first-line antidepressants. by using the active isomer only. 1997) precipitating many significant drug interactions. 104A new review of escitalopram can be found in Section 10. Fluvoxamine **Fluvoxamine was the first of the currently available SSRIs to be marketed in the UK.** Other mechanisms have been attributed to r-citalopram. 1982). which may account for some of the differences in efficacy seen between escitalopram and citalopram (Mork et al. The observation that escitalopram 10 mg is as effective as citalopram 20 mg confirms that escitalopram is responsible for most or perhaps the entire antidepressant efficacy of citalopram (Waugh & Goa. It is available as a generic preparation. Fluoxetine Fluoxetine is associated with a lower incidence of nausea than fluvoxamine but a higher incidence of rash. Citalopram Until the introduction of escitalopram. 1994) and it is pro-convulsant at high dose (Boeck et al. With respect to SSRI potency. 2001). although these are not firmly established. as outlined below. 2003). Sertraline is less problematic although enzyme inhibition is dose-related. sedation and sexual dysfunction than other SSRIs and more problems on withdrawal (Anderson & Edwards. It is available as a generic preparation. It has a long half-life. 337 . citalopram was the most serotonin selective of the SSRIs. Paroxetine Paroxetine is associated with a higher incidence of sweating. 2003).9. The issue of its safety in overdose is discussed below (see Section 11.

338 . 106Study IDs in title case refer to studies included in the previous guideline. Sertraline Sertraline is a well-tolerated SSRI.8 on antidepressant discontinuation symptoms). In two studies participants had depression with additional atypical features. 1991. In addition.5. It is available as a generic preparation. Thirty-three of these were included and 26 excluded.. a separate review107 of this drug was undertaken..106 The GDG used an existing review (Geddes et al.Pharmacological interventions 2001. A further two trials excluded by Geddes and colleagues (2002) were also considered part of the Feighner92 multicentre trial (Dunbar et al. Thus 107 trials are included in this review providing data from up to 11. All included studies were published between 1983 and 2003 and were between 4 and 24 weeks’ long (mean 6. 59 additional studies were identified from other reviews undertaken for this guideline. see also Section 11. In addition to the standard diagnostic criteria. more than 80% of participants were aged 65 years and over (although only eight of these reported extractable efficacy outcomes). Twenty-four studies were of inpatients. Feighner & Boyer. In the remaining 26. The Geddes and colleagues’ (2002) review included 126 studies of which 72 did not meet the inclusion criteria set by the GDG. including two identified from hand searching reference lists. References for these studies are in Appendix 18. or they were from mixed sources. Information about each study along with an assessment of methodological quality is in Appendix 17c.4. 2002) as the basis of this section. 1994). one trial (Peselow et al.. 10.5 weeks). most studies required a minimum baseline HRSD score of between 10 and 22 on the 17-item version (61 studies) or 105Details of standard search strings used in all searches are in Appendix 8. 107The review on escitalopram was updated for this guideline – see Section 10. 51 of outpatients and six undertaken in primary care. it was either not clear from where participants were sourced. It is available as a generic preparation. for which the authors made their data available to the NCCMH team. Since this review did not cover escitalopram which achieved its UK licence in late 2001. which also contains a list of excluded studies with reasons for exclusions. Since the Geddes and colleagues’ (2002) review compared SSRIs with TCAs only.2 Studies considered for review of selective serotonin reuptake inhibitors apart from escitalopram105. 1989). It is more likely to be associated with upwards dosage titration during treatment than the other SSRIs (Gregor et al.. Another trial (Feighner1989) excluded by Geddes and colleagues (2002) was included in this review because it contained tolerability data (which Geddes and colleagues [2002] did not include). A total of 97 trials were excluded.442 participants. In 11 studies. 1989) included in the original review was considered to be part of a multicentre trial (Feighner92) rather than a separate trial.

phenelzine. There is evidence suggesting that there is a statistically significant difference favouring third-generation110 antidepressants over SSRIs on reducing symptoms of depression as measured by the HRSD or MADRS. nortriptyline.05. N 3.03 to 0. 0. amitriptyline. 109Studies 339 . 108The forest plots can be found in Appendix 19c. –0.19).2) ● SSRIs and TCAs (K 15. where 50% of participants left treatment early were retained in the analysis since removing them made no difference to the results. whether combined as a group or divided by drug class: There is evidence suggesting that there is a statistically significant difference favouring other antidepressants over SSRIs on reducing symptoms of depression as measured by the HRSD or MADRS. N 4. clomipramine. fluvoxamine. N 469.02 to 0. Tolerability data.3 Clinical evidence statements for selective serotonin reuptake inhibitors apart from escitalopram108 Effect of treatment on efficacy There is no clinically important difference between SSRIs and other antidepressants. apart from third-generation antidepressants: There is evidence suggesting that there is no clinically important difference on reducing symptoms of depression in inpatients as measured by the HRSD or MADRS between: ● SSRIs and other antidepressants (K 20.13. venlafaxine and reboxetine. 95% CI. doxepin. SMD 0.01 to 0. maprotiline. –0.665. SMD 0. There is evidence suggesting that there is no clinically important difference on reducing symptoms of depression as measured by the HRSD or MADRS between: ● SSRIs and TCAs (K 49.09. SMD 0. SMD 0. 95% CI. paroxetine and sertraline) with amineptine.Pharmacological interventions between 18 and 22 on the 21-item version (28 studies). desipramine.24). trazodone. 0. imipramine. SMD 0. N 970. The ten studies reporting MADRS scores required minimum baseline scores of between 18 and 30.22). but the size of this difference is unlikely to be of clinical importance (K 17. mianserin. dothiepin/dosulepin. on which this review is based and for which the data were made available to the GDG.08. –0. 10.668.4. have been extracted by the NCCMH team. mirtazapine. 95% CI.06 to 0. 95% CI.073. 110Mirtazapine. lofepramine. included only one outcome measure (mean endpoint scores) and did not include tolerability data.03. 95% CI. N 1258. moclobemide. SMD 0.01 to 0. Data were available to compare SSRIs (citalopram. but the size of this difference is unlikely to be of clinical importance (K 82109. 95% CI.12). venlafaxine and reboxetine.15 to 0. N 8.12) ● SSRIs and MAOIs (K 7. Effect of setting on treatment efficacy In inpatients there is no difference between the efficacy of SSRIs and other antidepressants. fluoxetine. The Geddes and colleagues’ (2002) review.12. but not additional efficacy outcomes. –0.

05 to 0. There is evidence suggesting that there is a statistically significant difference favouring SSRIs over alternative antidepressants on reducing the likelihood of patients reporting adverse effects. SMD 0. There is insufficient evidence to determine whether there is a clinically important difference between SSRIs and MAOIs on reducing symptoms of depression as measured by the HRSD or MADRS in outpatients.442.85).96). 0. SMD 0. There is evidence suggesting that there is no clinically important difference on reducing symptoms of depression as measured by the HRSD or MADRS in outpatients between SSRIs and TCAs (K 24. There is strong evidence suggesting that there is a clinically important difference favouring SSRIs over alternative antidepressants on reducing the likelihood of patients leaving treatment early due to side effects (K 89. SMD 0. N 4666. 0 to 0. but the size of this difference is unlikely to be of clinical importance (K 38. Acceptability and tolerability of treatment There is evidence suggesting that there is a statistically significant difference favouring SSRIs over alternative antidepressants on reducing the likelihood of patients leaving treatment early for any reason. 95% CI. but the size of this difference is unlikely to be of clinical importance (K 9. 95% CI. SMD 0. In outpatients there is no difference between the efficacy of SSRIs and other antidepressants: There is evidence suggesting that there is a statistically significant difference favouring other antidepressants over SSRIs on reducing symptoms of depression as measured by the HRSD or MADRS in outpatients. N 10898.07).71 to 0. 0. but the size of this difference is unlikely to be of clinical importance (K 97. 95% CI. N 5658. 340 .096. 95% CI. There is evidence suggesting that there is a statistically significant difference favouring ‘third-generation’ antidepressants over SSRIs on reducing symptoms of depression as measured by the HRSD or MADRS in outpatients. N 2304. N 2. RR 0. N 67.12).06. but the size of this difference is unlikely to be of clinical importance (K 42.87 to 0. 95% CI.91 to 0. 0.21). There is a similar picture in primary care: There is evidence suggesting that there is no clinically important difference between SSRIs and other antidepressants on reducing symptoms of depression as measured by the HRSD or MADRS in primary care (K 4.22).94.02.78.08. RR 0.1). RR 0. N 11. SMD 0.91. 95% CI. –0. 95% CI.09 to 1. There is insufficient evidence to determine whether there is a clinically important difference between SSRIs and MAOIs on reducing symptoms of depression as measured by the HRSD or MADRS in inpatients. 0.07 to 0. 95% CI. 0. N 922.58.05 to 0.97).Pharmacological interventions There is some evidence suggesting that there is a clinically important difference favouring third-generation antidepressants over SSRIs on reducing symptoms of depression as measured by the HRSD or MADRS in inpatients (K 1. –0.13.

but the size of this difference is unlikely to be of clinical importance (K 17. but a relatively large number of studies (compared with the number previously available) have been published since then and so the review has been updated for this guideline.82 to 0. The marketing authorisation holder. both published and unpublished double-blind RCTs were sought that compared escitalopram either with placebo or with another antidepressant.81 to 0. N 6446.10).** 10.77).1 ESCITALOPRAM Introduction Escitalopram was reviewed in the previous guideline. RR 0.5 10. 341 . There is evidence suggesting that there is a statistically significant difference favouring SSRIs over TCAs on the likelihood of patients reporting adverse events.62 to 0. was also contacted for data. There is strong evidence suggesting that there is a clinically important difference favouring SSRIs over TCAs on reducing the likelihood of patients leaving treatment early due to side effects (K 59. 0.5. References for studies from the previous guideline are in Appendix 18. 10. due to their safety in overdose (see Section 11.69. 10.2 Databases searched and the inclusion/exclusion criteria Information about the databases searched for published trials and the inclusion/exclusion criteria used are presented in Table 68. 10. 95% CI. 0. 95% CI.9). N 6145.5. RR 0. For the present review. Some of the studies used in the previous 111Study IDs in title case refer to studies included in the previous guideline and study IDs in capital letters refer to studies found and included in this guideline update. Details of the search strings used are in Appendix 8.86. Lundbeck.3 Studies considered111 A total of six trials were included in the review in the previous guideline and these were supplemented by another 18 trials. RR 0.88. They may have an advantage for those with suicidal intent. 95% CI.93).4 Clinical summary of selective serotonin reuptake inhibitors apart from escitalopram SSRIs are relatively well-tolerated drugs with equal efficacy compared with alternative antidepressants.5.4. 0.Pharmacological interventions A sub-analysis against TCAs showed similar results: There is evidence suggesting that there is a statistically significant difference favouring SSRIs over TCAs on reducing the likelihood of patients leaving treatment early for any reason but the size of this difference is unlikely to be of clinical importance (K 62. N 1846.

ICD or similar criteria Escitalopram. 342 . EMBASE. and to ascertain effectiveness against individual drugs (in particular. Sub-analyses by dose indicated that both 10 and 20 mg doses were effective. more people left treatment early for any reason and because of side effects. Those that used a fixed dose of 10 or 20 mg were included in sub-analyses by dose. placebo. The full evidence profiles and associated forest plots can be found in Appendix 16c and Appendix 19c. therefore. although overall effect sizes were small and the quality of evidence graded moderate (largely because of heterogeneity). other antidepressants review that had been unpublished. PsycINFO. Table 71 gives the summary evidence table for escitalopram compared with all other antidepressants together. Escitalopram versus all other antidepressants Twenty one studies were found that compared escitalopram with other antidepressants. Sub-analyses were undertaken to assess the effect of the severity of depression at baseline and by dose. citalopram and other antidepressants are below). The full evidence profiles and associated forest plots can be found in Appendix 16c and Appendix 19c. although effect sizes were greater and graded moderate with the larger dose. citalopram).4 Clinical evidence Escitalopram versus placebo Eleven studies were found that compared escitalopram with placebo. respectively. other SSRIs and non-SSRI antidepressants. 10. The summary evidence profile can be found in Table 70. and with a range of other antidepressants. Summary study characteristics of the included studies are presented in Table 69 with full details in Appendix 17c. However. which also includes details of excluded studies. have since been published with different first authors.5.Pharmacological interventions Table 68: Databases searched and inclusion/exclusion criteria for clinical effectiveness of pharmacological treatments Electronic databases Date searched Update searches Study design Population Treatments MEDLINE. Escitalopram was effective when compared with placebo. (Separate analyses follow for escitalopram compared with SSRIs. CINAHL Database inception to January 2008 July 2008 RCT People with a diagnosis of depression according to DSM. Five studies in the current review are unpublished and supplied by the drug’s manufacturer. Data were available to compare escitalopram with placebo. and more people taking 20 mg reported side effects compared with those taking 10 mg. the study identifier has changed for some studies.

Table 69: Summary study characteristics of escitalopram Versus citalopram 6 RCTs (1917) (1) Burke2002* (2) COLONNA2005 (3) LEPOLA2003† (originally Montgomery2001) (4) MOORE2005 (5) SCT-MD-02† (6) YEVTUSHENKO 2007† (1) BALDWIN2006D (2) BOULENGER2006 (3) KASPER2005† (4) MAO2008 (5) SCT-MD-09 (6) SCT-MD-16 (7) SCT-MD-27† (8) VENTURA2007 (originally Alexopoulos2003) 8 RCTs (2086) 7 RCTs (2191) (1) BIELSKI2004 (originally Bielski2003) (2) CLAYTON2006C study 1† (3) CLAYTON2006C study 2† (4) KHAN2007B (5) Montgomery2002 (6) NIERENBERG 2007B† (7) WADE2007 Versus other SSRIs Versus other antidepressants Versus placebo No. trials (total participants) 11 RCTs (3409) Study IDs (1) BOSE2008 (2) Burke2002* (3) CLAYTON2006C study 1† (4) CLAYTON2006C study 2† (5) KASPER2005† (6) LEPOLA2003† (originally Montgomery2001) (7) NIERENBERG 2007B† (8) SCT-MD-02† (9) SCT-MD-26 (10) SCT-MD-27† (11) Wade2002 (1) 369/64 (2) 357/74 (3) 310/72 (4) 294/67 (1) 325/73 (2) 459/68 (3) 338/76 (4) 240/56 Pharmacological interventions N/% female (1) 267/58 (2) 366/64 (3) 283/61 (4) 286/54 (1) 198/58 (2) 284/61 (3) 297/54 (4) 278/60 Continued 343 .

9 mg (2) 20 mg (3)–(4) 10 mg (5)–(6) 20 mg (7) 16.6 mg (11) 10 mg . if given) (1) 20 mg (2) 10 mg and 20 mg (3)–(4) 13 mg (5)–(7) 10 mg (8)–(9) 10–20 mg (10) 16.Table 69: (Continued) Versus citalopram (5) 257/52 (6) 330/57 (5) 30/87 (6) 205/62 (7) 274/55 (8) 215/66 (5) 293/71 (6) 547/65 (7) 294/72 Versus other SSRIs Versus other antidepressants 344 (1) 40 (2) 46 (3) 43 (4) 45 (5) 42 (6) 35 (1) 45 (2) 44 (3) 75 (4) 39 (5) 39 (6) 37 (7) 40 (8) 39 (1) 37 (2) 36 (3) 37 (4) 42 (5) 48 (6) 42 (7) 44 (1) 10 mg and 20 mg (2)–(3) 10 mg (4) 20 mg (5) 10–20 mg (6) 10 mg (1) 13.1 mg (6) 10 mg (7) 20 mg Pharmacological interventions Versus placebo (5) 354/76 (6) 310/72 (7) 411/65 (8) 258/52 (9) 309/59 (10) 271/55 (11) 294/NA Mean age (1) 68 (2) 40 (3) 36 (4) 37 (5) 75 (6) 43 (7)–(8) 42 (9) 39 (10) 40 (11) 41 Escitalopram dose (mean.6 mg (8) 10 mg (1) 20 mg (2)–(3) 13 mg (4) 10/20 mg (5) 12.

3 mg (2) Paroxetine 20–40 mg (3)–(4) Fluoxetine 20 mg (5)–(6) Fluoxetine 40 mg (7) Sertraline 113.6-month continuation (8) 8 (9) 14 (10)–(11) 8 Pharmacological interventions *4-armed trial.1 mg (8) Sertraline 50–200 mg (1) Venlafaxine extended release (XR) 225 mg (2) Bupropion XL 323 mg (3) Bupropion XL 309 mg (4) Duloxetine 60 mg (5) Venlafaxine-XR 95. 345 .6-month continuation (7) 24 Setting (1)–(2) Outpatients (3)–(4) Unclear (5) Primary care and specialist (6) Primary care (7)–(10) Outpatients (11) Primary care (1) 8 (2) 6 months (3)–(5) 8 (6) 6 (1) 8 (2) 6 months (3)–(8) 8 (1)–(2) Outpatients (3) Primary care (4)–(6) Outpatients (1) Primary care (2) Outpatients (3) Primary care and specialist (4) Outpatients and inpatients (5)–(8) Outpatients Length of treatment (weeks) (1) 12 days (2)–(6) 8 (7) 8. †3-armed trial. if given) (1)–(11) Placebo (1) Citalopram 40 mg (2) 20 mg (3) 20/40 mg (4) 40 mg (5) 20–40 mg (6) 10 mg and 20 mg (1) Paroxetine 26.2 mg (6)–(7) Duloxetine 60 mg (1)–(3) Unclear (4) Outpatients (5) Primary care (6) Outpatients (7) Outpatients and primary care (1)–(5) 8 (6) 8.Comparator (mean dose.

35 to –0.88 (0.3%) Leaving treatment Leaving treatment Number reportearly early due to side ing side effects effects RR 1.23 (–0.1 versus 58.1 versus 68.26 (–0.6%) Moderate 3495 Pharm Esc 01.29) (22 versus 19.1 versus 65.02 RR 0.72 to 0.34 to –0.54) (5.75 to 1. n Forest plot number Pharm Esc 01.98) (53. n 2930 High 11.94 to 1.04 SMD –0.92 (0.95 to 1.7%) High K 8.11 (0.9 versus 2.15) (61.8 (1.346 Non-remission Mean depression scores at endpoint/mean change SMD –0.75 to 0.46 to –0.15) K K 6.99 (0.3) (19.9 versus 18. n 1821 K 10.4%) K 9.2%) Quality of evidence Moderate Number of studies/ participants K 11.7 versus 64.02 (0.73) (6.2%) Moderate K 11.05 RR 0.06 RR 2.18 to 2.09 (1.3 versus 3.01) SMD –0.9%) RR 1.07 RR 1.9%) RR 1.82 to 0. n 3495 Pharm Esc 01.15) (71.41 to –0.28 (–0.7%) RR 0.06) (62.81 to 1.94) (61.81 (0.9 to 4. n 3456 Pharm Esc 01.04 to 1.13)/ SMD –0.04 (0.24 (–0.88) (49.01 10 mg effect size RR 0.03 Pharm Esc 01.8%) . n Table 70: Summary evidence profile for escitalopram versus placebo Pharmacological interventions Non-response All data: effect size RR 0.8 versus 60.84 (0. n 2490 Pharm Esc 01.7 versus 61.19) Moderate High 2871 Pharm Esc 01.

1386 K n 4.07 Quality of evidence Moderate Number of studies/ participants K 1.5%) (85. 1386 K n 3.39) (10.47) (1.46 (–0. 1386 Forest plot number Pharm Esc 01. n 3. 1145 K K 3.74 to –0.06 RR 1.01 Pharm Esc 01.03 Pharm Esc 01.06 3.06 to 1.Quality of evidence Moderate K n Pharm Esc 01.17 (0.77) (28.77 to 1.6%) Pharm Esc 01.01 Pharm Esc 01.21 (1. 974 Moderate High Moderate Low Low High Number of studies/ participants K n 4.6 versus 70. n 247 K Low Moderate 1.02 Not reported SMD –0.48 (–0.6 versus 73%) – – – RR 4.68 (0. n 964 1025 K n 4.84) (49.07 20 mg effect size RR 0.24 to 14.71 to –0.5%) Moderate K 1. n 247 Pharm Esc 01.05 Pharmacological interventions 347 .8 versus 24. n 242 K 1.05 Pharm Esc 01. n 247 Forest plot number Pharm Esc 01.23 RR 1.22) Moderate Moderate K Pharm Esc 01.2) SMD –0.4 versus 2.04 1. n 247 Pharm Esc 01.04 Pharm Esc 01.03 Pharm Esc 01.55 to 0.

64 RR 0.88 to 0.01 .9 (0.91 to 0.6 versus 8.85 to 0.06 RR 0.98) (5. n 19.07 (–0.02 17.98) (18. 5206 RR 0.53 to 0.07 Table 71: Summary evidence profile for escitalopram versus all other antidepressants Pharmacological interventions Non-response Effect size RR 0.17 to –0.94 (0.348 Non-remission Mean depression Leaving scores at treatment endpoint/mean early change SMD –0.96) (37. n 3009 K 5158 n 21.78) (0.9 versus 21.05 Moderate High K n 20. 5832 Forest plot number Pharm Esc 02.93 (0.7 versus 41.98) (46.6%) Leaving treatment early due to side effects Number reporting side effects RR 0.03 Pharm Esc 02. 5807 Pharm Esc 02.74 to 0.85 (0.4%) Quality of evidence High Number of studies/ participants K n 19.12 to –0. 6192 Pharm Esc 02.1 (–0.9 versus 64.04 11. 4839 Pharm Esc 02.02) SMD –0.4%) High K n 17.6%) (63.3 versus 49.02) High High K K Pharm Esc 02.7%) High K n Pharm Esc 02.

The full evidence profiles and associated forest plots can be found in Appendix 16c and Appendix 19c. respectively. there were no clinically important differences on efficacy outcomes other than compared with citalopram. although the numbers reporting side effects were roughly equal. The summary evidence profile can be found in Table 73. escitalopram is more effective although the effect sizes are small. Escitalopram versus non-selective serotonin reuptake inhibitors Seven studies were found that compared escitalopram with non-SSRI antidepressants.Pharmacological interventions Compared with all antidepressants for which there are data. These can be found in Appendix 19c (forest plot numbers Pharm Esc 05 to Esc 11).5 Clinical summary Escitalopram is superior to placebo in the treatment of depression. Escitalopram is also compared with citalopram separately. This was particularly the case for escitalopram at 20 mg. with statistically significant differences versus SSRIs (although effect sizes are small and unlikely to be clinically important). which helped inform interpretation of the data. Compared with all SSRIs together. respectively. apart from compared with sertraline. although differences were small. where escitalopram was more effective with a small effect size. Several more detailed comparisons were considered by the GDG. Escitalopram is more effective than citalopram although the effect size is small.5. differences were again small and unlikely to be clinically important. It was also marginally more acceptable and tolerable. The full evidence profiles and associated forest plots can be found in Appendix 16c and Appendix 19c. escitalopram was more effective although effect sizes were small. Effect sizes compared with citalopram were largest. apart from sertraline. except against sertraline. Escitalopram was also more acceptable and tolerable than SSRIs. 349 . There were no clinically important differences between escitalopram and duloxetine. Escitalopram was mostly more acceptable and tolerable. but not against other antidepressants (duloxetine. Escitalopram versus selective serotonin reuptake inhibitors Eight studies were found that compared escitalopram with SSRIs. in addition to those presented above. although these were still relatively small. Compared with individual SSRIs. Escitalopram was more effective than other antidepressants. venlafaxine or bupropion on efficacy measures. It is at least as effective as other SSRIs and marginally better tolerated. venlafaxine and bupropion). Fewer participants taking escitalopram left treatment early for any reason or because of side effects compared with those taking other antidepressants. although all effect sizes favoured escitalopram. The summary evidence profile can be found in Table 72. 10. There is some evidence that 20 mg may be more effective than 10 mg. However. but at the expense of increased side effects. although differences were again small.

7%) High RR 0. n 2434 13.29) (5.11) (15.06 SMD –0.6%) Leaving treatment early Leaving Number reporttreatment early ing side effects due to side effects RR 0.89 (0.1%) High K Pharm Esc 03.97) (36.82 to 0. n Moderate 4010 High K 13.7 versus 64.2 versus 45.6 versus 6.6%) Quality of evidence High .6 to 1.6%) Quality of evidence High Number of studies/ participants K 12.04 RR 0.94 (0.75 (0.95) (40.94) (41 versus 50.12 (–0.2%) High 3650 Pharm Esc 03.28 to –0.4%) Low RR 0.8 versus 18.8 (0.03) SMD –0.8 versus 67.86 (0.1 (–0.58 to 0.82 (0.02 Citalopram effect size RR 0.02) High High 10.11 (–0.96) (5.98) (41.7%) High K 13.2%) High RR 0.350 Non-remission Mean depression scores at endpoint/mean change SMD –0.1 versus 39.9 (0.89 to 1.2 versus 15.49 to 1.98) (64. n 3652 Pharm Esc 03.05) High High 3024 K K 9.07 Table 72: Summary evidence profile for escitalopram versus SSRIs Pharmacological interventions Non-response All SSRIs effect size RR 0.19 to –0.02) (64.71 to 1. n 3337 K 14.95 (0.8 versus 7. n 3639 Pharm Esc 03.72 to 0.08 RR 0.18 to –0.10 RR 0.82 (0.6 versus 46.05 Pharm Esc 03.17 (–0.83 to 0.03) (16. n Forest plot number Pharm Esc 03. n Pharm Esc 03.6%) RR 0.76 to 0.85 (0.24 to 0) SMD –0.9 to 0.

8 to 1.4%) RR 0.24 to 0. n 813 K 4.06 Pharm Esc 03. n 968 K K 4.9%) RR 0.81 to 1.25) SMD 0.87 to 1) (83.2 versus 89%) High K 2. n 783 K n 3. n 483 Quality of evidence High Number of studies/ K participants 2. n 483 SMD –0.06) SMD –0.05 Pharm Esc 03.04 RR 0.02 Fluoxetine effect size RR 0.2 versus 3.6 versus 8.17 to 0.98 versus 21. n Moderate High 211 477 Low K 2. n 489 K K 1.07 3.22) (50.7%) High K 2.6%) High K 2.86 to 1. n 2.29 to 0.8 versus 35.7%) High K Pharm Esc 03. n 1143 1639 K 6.10 5.22) (4.92 (0.Number of studies/ K participants Pharm Esc 03.47 to 1.08 RR 1. n 1583 Forest plot number Pharm Esc 03.08) (39.02 (0.26) (6.42) (19.11 (0.9 versus 48. n High High Low High 805 Pharm Esc 03.01 (0.3 versus 16.93 (0.06 (–0. n 6.08 Pharm Esc 03.28) (35. n 1569 K 5.6%) Pharm Esc 03.9%) Quality of evidence High Number of studies/ participants K 3.19) RR 1.05 Pharm Esc 03.3%) Pharm Esc 03.34 to –0.3 versus 61.13) RR 0.06 Pharm Esc 03.92 (0.06) (44.07 Pharm Esc 03. n 489 Pharmacological interventions Continued 351 .02 (–0.7%) High K 4.04 RR 1. n 489 SMD –0.58 to 1.03) (56.8 to 1. n 783 Forest plot number Pharm Esc 03.74) (19.2 (–0.19 (0.91 (0.77 (0.01 (–0.02 Sertraline effect size RR 1. n 1594 K 3.92 (0.78 to 1. n 804 Pharm Esc 03.10 RR 0.82 to 1.8 versus 35.38 to 3.6 versus 49. n 1924 K 5. n 449 759 K 4.

65 (0. n 321 772 K High 2.5%) Low K Pharm Esc 03.9%) Quality of evidence Low Number of studies/ participants K 2.10 Pharmacological interventions Non-response Forest plot number Pharm Esc 03.04) (6.9%) Pharm Esc 03.04 RR 0.3 versus 10.49 to 0.08 High K 2.85) (16.92 (0.10 Pharm Esc 03.6%) (62.38 to 0. n 782 Pharm Esc 03.05 Pharm Esc 03.11) (33.07 Pharm Esc 03. n 2.06 SMD 0.04 RR 0.352 Table 72: (Continued) Non-remission Mean depression scores at endpoint/mean change Pharm Esc 03.76 to 1. n 782 Pharm Esc 03.65 RR 0. n Pharm Esc 03.9 versus 26.05 Pharm Esc 03.07 RR 0.85 to 1.08 Leaving treatment early Leaving Number reporttreatment early ing side effects due to side effects Pharm Esc 03.17) (24.36) (0.06 784 K K 1.02 Paroxetine effect size RR 0. n 784 Pharm Esc 03. n 784 Forest plot number Pharm Esc 03.11 to 0.31 to 1.27) Moderate Very low 2.33) SMD –0.7 versus 36.7 versus 67%) Low K 2.06 (–0.02 .92 (0.1 versus 24.94 (0.73 to 1.11 (–0.

1 versus 36.88 (0.78 (0.01 . n RR 0.88 (0. n 557 Pharm Esc 04. n 2.08 (–0.05 K K 1.94 (0. n 1120 Forest plot number Pharm Esc 04.05 RR 0.03 (–0.17 to 1.3%) High K Pharm Esc 04.47 (0.4%) Pharm Esc 04. n Pharm Esc 04.13) (55.07 Not reported RR 0.06 RR 0.94 to 1.86 (0.01 Pharm Esc 04. n 491 Forest plot number Pharm Esc 04.22 to 0.63 to 1.31) (6.57 to 1.2%) Leaving treatment early Leaving treatment early due to side effects Number reporting side effects Non-response Duloxetine effect size RR 0. n 288 483 K 2. n 2.5 versus 13.03 Pharm Esc 04.49 to 1) (21.05 Bupropion XL effect size RR 0.07) (39.15 to 0.08 (0.3 versus 29.3%) Low K 2.37 to 0.19 (–0.21) (58.6 versus 56%) Moderate K Pharm Esc 04. n 571 Pharm Esc 04.01 Pharm Esc 04.06 High K 1.4 versus 11.98 (0.6%) Moderate K n 2.29) Moderate Low 2.1) (66.16 to 3. n 1120 K 3.98 (0.1%) Venlafaxine effect size RR 0.47 (0.8%) Quality of evidence Low Number of studies/ participants K 3. n 572 Pharm Esc 04.11 to 0.9%) RR 0.15) (43.05 (–0.02 (0.8%) Quality of evidence Moderate – – – Pharmacological interventions 353 Pharm Esc 04.3 versus 6.83 to 1.72 to 1.79 to 1.Table 73: Summary evidence profile for escitalopram versus non-SSRIs Non-remission Mean depression scores at endpoint/ mean change SMD –0.89) (5.02 RR 0.68 to 1.8 versus 45.81 (0. n 287 809 K 3.23) (19.4 versus 59. n Moderate 491 RR 0.09) (32.04 Not reported SMD –0.06 RR 1.97 (0.2%) Very low K 2.04) SMD 0. n 1120 Moderate Moderate High K 2.12) Not reported High Not reported K 2. n 293 Pharm Esc 04.02 RR 0.4 versus 48.04 (–0. 571 Forest plot number Pharm Esc 04.8 versus 77. 571 3.32) SMD –0.22) (40.9 versus 41. n 529 Pharm Esc 04 491 K K 1.2%) RR 1.8 versus 25.42 to 0.02 Number of studies/ participants K n 2.81 to 1.17) Moderate High 1120 Pharm Esc 04.7) (4.7 (0.82 to 1.7%) Moderate K 2.11) (78.41) (27.78 to 1.7%) Quality of evidence Moderate Number of studies/ participants K 2.25 to 0.2 versus 70.03/04 SMD 0. n 491 Pharm Esc 04.07 RR 0.9 versus 22.

20 points) difference between the two arms at 12 weeks. The primary outcome reported was the HDRS. switched SSRIs. In total.and patient-rated measures of depressive symptoms were taken at 12 and 26 weeks.Pharmacological interventions Overall. are available. although the cost per QALY was towards the upper end of the accepted NICE range of £20. Patients had to meet a minimum criterion score of 12 on the HDRS and symptoms had to have persisted for at least 8 weeks. if thought necessary.6 THE THREAD STUDY The THREAD study (Kendrick et al. the lack of a placebo control. In total. Significant differences were also identified in remission and response rates and a cost-effective analysis suggested that the addition of SSRIs to supportive care might be cost effective. The study had a number of limitations including the open label design. particularly when other treatments with potentially greater acceptability to patients. open label. multi-centre RCT comparing SSRIs plus supportive care with supportive care alone for mild to moderate depression in primary care. that is.000/QALY. given the small effect size this study does not suggest changes to the recommendation from the original guideline that SSRIs should not be offered routinely in primary care for people with mild to moderate depression. they were discouraged to do so but GPs could also prescribe antidepressants in the supportive arm of the trial. 2009) is a pragmatic. Nevertheless it suggests that SSRIs could be of value in mild to moderate depression for people whose symptoms have persisted for some time. 4. SSRIs might be considered for patients with mild to moderate depression who have persistent symptoms. GPs prescribed and.. 8 and 12 weeks after the baseline assessment. Supportive care from GPs consisted of follow-up consultations 2. 10. However. This conclusion is broadly in line with the recommendation developed in the original guideline based on the review of the SSRIs. the quality of the evidence tended to be downgraded because of heterogeneity between trials. 10. It did not have a placebo arm and was close to real-life practice. such as a range of low-intensity psychosocial interventions.7 MONOAMINE OXIDASE INHIBITORS The following sections on MAOIs marked by asterisks (**_**) are from the previous guideline and have not been updated except for style and minor clarification. which showed a small (2. It was designed to address the question of the effectiveness of antidepressants in people with mild to moderate depression because of the uncertainty about the risk–benefit ratio in this group. the overall small effect size and the absence of effect on the patientrated BDI. 87% of patients in the SSRIs plus supportive care arm and 20% in the supportive care alone arm received SSRIs. Since escitalopram is still in patent its acquisition costs are relatively high compared with antidepressants available in generic form. no significant difference was identified on the BDI. 354 . 220 patients were recruited to the trial and outcomes on clinician. which was statistically significant. although it did improve other patient-rated measures.

113Study IDs in title case refer to studies included in the previous guideline. For this class of drugs.2 Moclobemide Introduction Moclobemide is a reversible selective inhibitor of monoamine oxidase A (a RIMA). It is not widely prescribed in the UK.1 Introduction **MAOIs exert their therapeutic effect by binding irreversibly to monoamine oxidase. Moclobemide is the only RIMA licensed in the UK. Information about each study along with an assessment of methodological quality is in Appendix 17c.7. which also contains a list of excluded studies with reasons for exclusions. MAOIs are still widely cited as being the most effective antidepressants for the treatment of atypical depression (see Section 11.5). Moclobemide is generally well-tolerated as it is associated with a low potential for producing anticholinergic side effects. Twenty-seven additional 112Details of standard search strings used in all searches are in Appendix 8. It has the advantages over the traditional MAOIs that strict dietary restrictions are not required. The first antidepressant drug synthesised was an irreversible MAOI and drugs in this class have been available in the UK for nearly 50 years. All MAOIs have the potential to induce hypertensive crisis if foods containing tyramine (which is also metabolised by monoamine oxidase) are eaten (Merriman. even in hospitalised patients. 1996). Reversible inhibitors of monoamine oxidase (RIMAs) have a much lower likelihood of causing a hypertensive crisis and dietary restrictions are usually not required. the enzyme responsible for the degeneration of monoamine neurotransmitters such as NA and serotonin. 355 . as opposed to the traditional MAOIs that irreversibly inhibit both monoamine oxidase A and monoamine oxidase B. potentially serious drug interactions and the availability of safer antidepressants have led to the irreversible MAOIs being infrequently prescribed in the UK. This results in increased monoamine neurotransmission. 1999) or drugs that increase monoamine neurotransmission are co-prescribed (Livingstone & Livingstone.7.Pharmacological interventions 10. 10. References for these studies are in Appendix 18. These foods and drugs must be avoided for at least 14 days after discontinuing MAOIs. However. drug interactions leading to hypertensive crisis are less problematic and shorter washout periods are required when switching to other antidepressants. weight gain and symptomatic postural hypotension. Dietary restrictions. the GDG chose to review phenelzine and moclobemide. Studies considered112.113 Forty-four studies were found in a search of electronic databases with 12 meeting the inclusion criteria set by the GDG and 32 being excluded.

Reynaert1995. Hebenstreit90. participants were classified as inpatients. Newburn1990. A total of 45 studies were excluded. Jouvent1998. N 111. Bougerol1992.13 to –0. Koczkas1989. Duarte1996. fluvoxamine and placebo. Random effects RR 0. and in ten.05).88. either a mixture of inpatients and outpatients or the setting was unclear. Casacchia1984. Sixteen studies compared moclobemide with TCAs (Bakish1992. Beaumont1993. Versiani1989. Tanghe1997. There is some evidence suggesting that there is a clinically important difference favouring moclobemide over placebo on increasing the likelihood of achieving at least a 50% reduction in symptoms of depression as measured by the HRSD (K 3. Silverstone94.149 participants. Clinical evidence statements for moclobemide compared with placebo114 Effect of treatment on efficacy outcomes There is some evidence suggesting that there is a clinically important difference favouring moclobemide over placebo on reducing symptoms of depression by the end of treatment as measured by the HRSD (K 3. Nair1995. 14 of which met inclusion criteria with 13 being excluded. Williams1993) and seven with placebo (Bakish1992. Random effects SMD –0. Larsen1989. KraghSorensen95. Lecrubier1995. as outpatients. 114The forest plots can be found in Appendix 19c.7. A total of 26 studies are therefore included in this review (Bakish1992. Lapierre1997.742 participants and tolerability data from up to 2. 356 .73 to 1. –1. Gattaz1995. Barrelet1991. In one study (Nair1995). 95% CI. Larsen1989. Silverstone94. Koczkas1989. Tanghe1997. Hell1994. Geerts1994. Williams1993) providing efficacy data from up to 1. Reynaert1995. All included studies were published between 1984 and 1998 and were between 4 and 7 weeks’ long (mean 5. fluoxetine. There is insufficient evidence to determine whether there is a clinically important difference between moclobemide and placebo on increasing the likelihood of achieving remission by the end of treatment as measured by the HRSD (K 2. Hebenstreit90. Guelfi1992. 0. Versiani1989). N 490. RR 0. in two. Silverstone1994. Versiani1989). Duarte1996. Casacchia1984. in a further seven studies. primary care. None of the included studies described participants as having depression with atypical features. Nair1995. Bougerol1992. Newburn1990. dothiepin/dosulepin. Geerts1994. In seven studies. eight with SSRIs (Barrelet1991.34 weeks). nortriptyline.6. Beckers1990. imipramine. Nair1995. Jouvent1998. Gattaz1995. Lecrubier1995. Larsen1989.5 to 0. Ose1992. the patients were exclusively older adults (aged 60 to 90 years). KraghSorensen95. Ose1992. Beckers1990. clomipramine. N 606. Data were available to compare moclobemide with amitriptyline.Pharmacological interventions studies were identified from other searches undertaken for this guideline. 95% CI. Beaumont1993. Participants received between 150 and 600 mg of moclobemide with most receiving at least 300 mg.99). Hell1994. Lapierre1997.07). Guelfi1992. 95% CI. 0.

Acceptability and tolerability of treatment There is evidence suggesting that there is no clinically important difference between moclobemide and other antidepressants on reducing the likelihood of leaving treatment early for any reason (K 20. 95% CI. N 785.85. this did not affect the results.79 to 1. 95% CI.32).79 to 0. Similar results were found in sub-analyses by antidepressant class and setting. There is evidence suggesting that there is a statistically significant difference favouring moclobemide over other antidepressants on reducing the likelihood of patients reporting side effects.02.12. Tanghe1997) were removed from this analysis to remove heterogeneity from the dataset.97. N 2070.04) ● reducing the likelihood of patients reporting side effects (K 5. RR 1. Random effects RR 0.12 to 0. 0.9.86 to 1. N 2458. 95% CI. RR 0. 116Two studies (Duarte1996.13).22) ● reducing the likelihood of leaving treatment early due to side effects (K 6. 95% CI. Random effects RR 1. Similar results were found in sub-analyses by setting but not by antidepressant class: There is evidence suggesting that there is no clinically important difference between moclobemide and SSRIs on reducing the likelihood of patients reporting side effects (K 6. N 402. Similar results were found in sub-analyses by antidepressant class and setting.92). 0. 0.75). RR 0. 95% CI. 95% CI.74 to 1. 0. There is strong evidence suggesting that there is a clinically important difference favouring moclobemide over other antidepressants on reducing the likelihood of leaving treatment due to side effects (K 18.93 to 1.11. Clinical evidence statements for moclobemide compared with other antidepressants115 Effect of treatment on efficacy outcomes There is evidence suggesting that there is no clinically important difference between moclobemide and other antidepressants on: ● reducing symptoms of depression by the end of treatment as measured by the HRSD (K 13116.6 to 2.94 to 1. 0. N 1222. N 1472.03). N 819.95.85 to 1. 95% CI.57. 115Ibid. SMD 0. RR 1. 95% CI. N 2292. 357 . N 519. 0.Pharmacological interventions Acceptability and tolerability of treatment There is insufficient evidence to determine if there is a clinically important difference between moclobemide and placebo on: ● reducing the likelihood of leaving treatment early for any reason (K 7. N 615. RR 1.11). RR 0.18) ● increasing the likelihood of achieving at least a 50% reduction in symptoms of depression by the end of treatment as measured by the HRSD or MADRS (K 13. 95% CI. 0. RR 0. 0. 95% CI.44 to 0. 0.11) ● increasing the likelihood of achieving remission by the end of treatment as measured by the HRSD (K 5. –0. but the size of this difference is unlikely to be of clinical importance (K 12.

Pharmacological interventions There is insufficient evidence to determine if there is a clinically important difference between moclobemide and SSRIs on reducing the likelihood of leaving treatment early due to side effects (K 6. There is strong evidence suggesting that there is a clinically important difference favouring moclobemide over TCAs on reducing the likelihood of leaving treatment due to side effects (K 12.3 Phenelzine Introduction Phenelzine is the best tolerated MAOI. 0. N 953.34 to 0. which also contains a list of excluded studies with reasons for exclusions. Swann1997. There is evidence that it is equally as effective as other antidepressants (TCAs and SSRIs). These provided efficacy data from up to 634 trial participants and tolerability data from up to 481 participants. 95% CI. Raft1981.7. 358 . RR 0. but insufficient evidence of its tolerability and acceptability. dizziness. 10.46. Established side effects include hypotension.64). N 1632. drowsiness.83. Quitkin1990119.96.56 weeks). 95% CI. 0. It has been associated with hepatotoxicity. RR 0. Georgotas86 was also the only study in which all participants were 55 years of age or older (mean 117Details of standard search strings used in all searches are in Appendix 8. Information about each study along with an assessment of methodological quality is in Appendix 17c. dry mouth and constipation. 0. 119The data from Quitkin1990 was supplied as raw individual patient data by the authors to the NCCMH review team. While moclobemide is equally as acceptable and tolerable to patients as SSRIs. Eight studies compared phenelzine with TCAs (Davidson81.118 Twenty-seven studies were found in a search of electronic databases with nine being included and 18 being excluded by the GDG. there is strong evidence that patients receiving moclobemide are less likely to leave treatment early due to side effects than patients receiving TCAs. Georgotas86.57). Davidson87.59 to 1. Robinson1983. Vallejo87) and one with SSRIs (Pande1996). Studies considered117. All included studies were published between 1981 and 1997 and were between 3 and 7 weeks’ long (mean 5. There is evidence suggesting that there is a statistically significant difference favouring moclobemide over TCAs on reducing the likelihood of patients reporting side effects but the size of this difference is unlikely to be of clinical importance (K 6. References for these studies are in Appendix 18. 95% CI. Clinical summary There is some evidence that moclobemide is more effective than placebo. N 660.76 to 0. RR 0.91). Participants were described as outpatients in eight studies and as inpatients in the other study (Georgotas86). 118Study IDs in title case refer to studies included in the previous guideline.

0.07.97. Data were available to compare phenelzine with amitriptyline. N 325. There is evidence suggesting that there is no clinically important difference between phenelzine and other antidepressants on reducing the likelihood of patients reporting adverse effects (K 1.02. N 285.09). N 385. A sub-analysis by antidepressant class gave similar results. 359 . RR 0. 120Not 121The licensed for use in the UK.28).83). imipramine. Clinical evidence statements for phenelzine121 Effect of treatment on efficacy outcomes There is some evidence suggesting that there is a clinically important difference favouring phenelzine over other antidepressants on increasing the likelihood of achieving a 50% reduction in symptoms of depression as measured by the HRSD (K 2.52 to 0. –0. 95% CI. 0.66. 95% CI. 95% CI.87 to 1.70). There is some evidence suggesting that there is a clinically important difference favouring phenelzine over TCAs on increasing the likelihood of achieving a 50% reduction in symptoms of depression as measured by the HRSD (K 1.83). There is evidence suggesting that there is no clinically important difference between phenelzine and other antidepressants on reducing symptoms of depression by the end of treatment as measured by the HRSD or MADRS (K 7.33 to 0. nortriptyline and fluoxetine. RR 0.97. Studies reported mean doses of between 30 and 90 mg of phenelzine. 0. N 634.52 to 0. –0. RR 0. N 60.66. 95% CI.40 to 0. N 594. 0. forest plots can be found in Appendix 19c. 95% CI. Acceptability and tolerability of treatment There is insufficient evidence to determine whether there is a clinically important difference between phenelzine and other antidepressants on reducing the likelihood of leaving treatment early for any reason and on reducing the likelihood of leaving treatment early due to side effects. Random effects SMD –0.55 to 1. There is evidence suggesting that there is no clinically important difference between phenelzine and TCAs on reducing symptoms of depression by the end of treatment as measured by the HRSD or MADRS (K 6. All participants in Pande1996 and 67% of those in Quitkin1990 were diagnosed with depression with additional atypical features. Random effects SMD –0. There is insufficient evidence to determine whether there is a clinically important difference between phenelzine and other antidepressants on increasing the likelihood of achieving remission by the end of treatment as measured by the HRSD (K 3.Pharmacological interventions age 65 years). There is insufficient evidence to determine whether there is a clinically important difference between phenelzine and SSRIs on any efficacy measure or between phenelzine and TCAs on reducing the likelihood of achieving remission by the end of treatment. Random effects RR 0. 95% CI. desipramine120.27).

(A separate review of the pharmacological treatment of atypical depression is provided in Section 11. Evidence from studies comparing phenelzine with SSRIs was too weak to draw any conclusions. one is licensed primarily for depression. reboxetine and venlafaxine.8.8. This is also evident in comparisons with TCAs alone. and the other for stress urinary incontinence. 10. The aim was to broaden the mechanism of action beyond serotonin in order to improve efficacy without incurring the side effects or toxicity in overdose associated with the TCAs. mirtazapine. Duloxetine is associated with nausea and headache.2 Duloxetine Introduction Duloxetine has been licensed since the publication of the previous guideline.8 THIRD-GENERATION ANTIDEPRESSANTS122 Sections on third-generation antidepressants marked by asterisks (**_**) are from the previous guideline and have not been updated except for style and minor clarification. These findings are probably explained by the high proportion of patients with depression with atypical features in the studies reporting response (71% of patients had depression with atypical features) and remission (56% of patients had depression with atypical features).Pharmacological interventions Clinical summary There is some evidence suggesting a superior efficacy for response for phenelzine compared with other antidepressants. 10. Duloxetine is available under two brand names from the same manufacturer. and is a weak inhibitor of dopamine reuptake.5. It is one of the few antidepressants that has been tested in double-blind.1 Introduction **This diverse group of antidepressants was marketed after the SSRIs. There is insufficient evidence to draw any conclusions on the comparative tolerability of phenelzine against alternative antidepressants. to avoid confusion with the guideline’s use of ‘other antidepressants’ to mean all other antidepressants. 10.** The following drugs are reviewed in this section: duloxetine (a new review for this updated guideline). placebo-controlled trials in elderly patients. and can also increase blood pressure. 360 . It is similar to venlafaxine in that it inhibits the reuptake of both serotonin and NA.) There is no difference in mean endpoint scores between the two groups of treatments in patients with depression regardless of additional atypical features. 122Although these are classified ‘other antidepressants’ by the BNF. the GDG uses the term ‘third-generation antidepressants’ to describe this group of drugs.

ICD or similar criteria Duloxetine. Eli Lilly. Summary study characteristics of the included studies are presented in Table 75 with full details in Appendix 17c. and with other antidepressants (SSRIs or venlafaxine). three trials continued treatment for those with at least a partial response ( 30% improvement in baseline depression scores). with duloxetine at different doses. Information about the databases searched for published trials and the inclusion/ exclusion criteria used are presented in Table 74.org from where full trial reports were downloaded. Section 12.) Only data from patients given at least the licensed dose (60 mg) were included in the analyses. PsycINFO. CINAHL Database inception to January 2008 July 2008. which also includes details of excluded studies. apart from in trials that used a variable dose and in trials where comparisons with the licensed dose were possible.3] because it re-randomised patients who did not respond to acute phase treatment. The marketing authorisation holder.Pharmacological interventions Table 74: Databases searched and inclusion/exclusion criteria for clinical effectiveness of pharmacological treatments Electronic databases Date searched Update searches Study design Population Treatments MEDLINE. both published and unpublished double-blind RCTs were sought that compared duloxetine either with placebo or with another antidepressant. (One trial is also included in the review of treatment-resistant depression [see Chapter 12. In addition. Eli Lilly. 18 trials (four unpublished) were included with nine excluded (seven unpublished). Details of the search strings used are in Appendix 8. 123Study IDs in title case refer to studies included in the previous guideline and study IDs in capital letters refer to studies found and included in this guideline update. placebo. References for studies from the previous guideline are in Appendix 18. was also contacted for data. January 2009 RCT People with a diagnosis of depression according to DSM. other antidepressants Databases searched and the inclusion/exclusion criteria For the present review.clinicaltrialresults. 27 acute-phase trials were sourced from searches of electronic databases and from the website of the drug’s manufacturer. Data were available to compare duloxetine with placebo. EMBASE. which included links to the clinical trials website www. Studies considered123 In total. In all. 361 .

362 Table 75: Summary study characteristics of studies of duloxetine Versus different doses 5 RCTs (1242) (1) DETKE2004† (2) ELI LILLY HMAT-A‡ (3) GOLDSTEIN2004† (4) PERAHIA2006B† (5) WHITMYER2007‡ 12 RCTs (3.069) Pharmacological interventions Study IDs (1) BRANNAN2005A (2) BRECHT2007 (3) DETKE2002 (4) DETKE2002A (5) DETKE2004† (6) ELI LILLY HMAQ‡ (7) ELI LILLY HMAT-A‡ (8) GOLDSTEIN2002‡ (9) GOLDSTEIN2004† (10) NIERENBERG2007B‡ (11) PERAHIA2006B† (12) RASKIN2007 N/% female (1) 282/65 (2) 327/74 (3) 267/69 (4) 245/67 (5) 281/73 (6) 157/67 (7) 174/62 (8) 140/64 .367) (1) DETKE2004† (2) ELI LILLY HMAQ‡ (3) ELI LILLY HMBU (4) ELI LILLY HMCQ‡ (5) ELI LILLY HMAT-A‡ (6) GOLDSTEIN2002‡ (7) GOLDSTEIN2004† (8) KHAN2007B (9) LEE2007 (10) NIERENBERG2007B‡ (11) PERAHIA2006B† (12) WADE2007 (1) 180/73 (2) 119/67 (3) 323/71 (4) 504/66 (5) 173/62 (6) 103/64 (7) 178/62 (8) 278/60 Versus other antidepressants (1) 188/73 (2) 175/62 (3) 177/62 (4) 196/70 (5) 506/64 Versus placebo No. trials (total participants) 12 RCTs (3.

80 mg (10) 60 mg (11) 80 mg. 120 mg (12) 60 mg Pharmacological interventions Continued 363 . 80 mg (8) 120 mg (9) 40 mg*.(9) 180/62 (10) 410/65 (11) 295/70 (12) 311/60 (1) 43 (2) 44 (3) 40 (4) 45 (5) 43 (1) 43 (2) 40 (3) 44 (4) 42 (5) 44 (6) 41 (7) 44 (8) 42 (9) 38 (10) 42 (11) 45 (12) 45 (1) 80 mg (2) 40–120 mg (3)–(4) 60 mg (5) 80 mg (6) 120 mg (7) 80 mg (8)–(10) 60 mg (11) 80 mg (12) 60 mg (9) 478/70 (10) 547/65 (11) 190/70 (12) 294/72 Mean age (1) 40 (2) 50 (3) 41 (4) 42 (5) 43 (6) 40 (7) 44 (8) 41 (9) 40 (10) 42 (11) 45 (12) 72 (1) 80 mg versus 120 mg (2)–(3) 40 mg versus 80 mg (4) 80 mg versus 120 mg (5) 30 mg versus 60 mg Duloxetine dose (1)–(4) 60 mg (5) 80 mg. 120 mg (6) 40–120 mg (7) 40 mg*.

75 mg (5) Paroxetine 20 mg (6) Fluoxetine 20 mg (7) Paroxetine 20 mg (8) Escitalopram 10 mg (9) Paroxetine 20 mg (10) Escitalopram 10 mg (11) Paroxetine 20 mg (12) Escitalopram 10 mg Outpatients (1) 9 weeks (2) 8 weeks (3)–(4) 12 weeks (5) 8 weeks (6) 10 weeks (7)–(12) 8 weeks (1) 6 months for partial responders (2) 6 months but data not available (11) 6 months for partial responders Versus other antidepressants 364 Outpatients (1)–(4) 8 weeks (5) 6 weeks (1) 6 months for partial responders (2) 6 months but data not available (4) 6 months for partial responders (5) Non-responders re-randomised (data in Chapter 12.3) Versus placebo Pharmacological interventions Comparator Placebo Setting Outpatients Length of treatment (1) 9 weeks (2) 8 weeks (3)–(4) 9 weeks (5) 8 weeks (6) 10 weeks (7)–(12) 8 weeks Continuation phase (length and inclusion criterion) (5) 6 months for partial responders (6) 6 months but data not available (11) 6 months for partial responders *Data not used as dose given less than licensed dose. Section 12. †4-armed trial.Table 75: (Continued) Versus different doses Duloxetine (doses as above) (1) Paroxetine 20 mg (2) Fluoxetine 20 mg (3) Venlafaxine 150 mg (4) Venlafaxine 150 mg. . ‡3-armed trial.

57. PERAHIA2006B). There was no difference in symptoms of depression or on acceptability and tolerability measures between duloxetine at either 80 or 120 mg and placebo. with only that for non-response approaching clinical importance. and 80 mg with 120 mg. although there were no extractable data in ELI LILLY HMAQ. There was little difference between the number of people receiving duloxetine who left treatment early for any reason and those receiving placebo on this measure. The numbers reporting side effects were high in both groups. Two trials specifically examined depression-related pain using the self-report Brief Pain Inventory (BPI) scale. Evidence from the important outcomes and overall quality of evidence are presented in Table 77. The full evidence profiles and associated forest plots can be found in Appendix 16c and Appendix 19c. Evidence from the important outcomes and overall quality of evidence are presented in Table 76. twice as many taking duloxetine as those taking placebo left specifically because of side effects while twice as many taking placebo left because of lack of efficacy. 30 mg with 60 mg. The data for duloxetine at different doses can be seen in the full evidence profiles and forest plots (Appendix 16c and Appendix 19c. there were similar effect sizes for duloxetine at different doses when these data were looked at separately. The full evidence profiles and associated forest plots can be found in Appendix 16c and Appendix 19c. There was an average reduction of three-quarters of a point (on an 11-point Likert scale) for the ‘average pain in last 24 hours’ item.Pharmacological interventions Clinical evidence Duloxetine versus placebo Although the effect sizes for all three efficacy outcomes for duloxetine (dose at least as large as the licensed dose of 60 mg) versus placebo were statistically significant and favoured duloxetine. respectively. Three studies continued patients who achieved at least partial response to acutephase treatment (defined as 30% decrease in baseline HAMD scores) (DETKE2004. of those leaving treatment early. largely because of the selective population included in the studies. respectively. although the effect sizes for duloxetine at 120 mg versus placebo was larger than those for lower does (WMD 2. although these data were of low quality largely because of heterogeneity. The quality of the evidence was moderate or low. Duloxetine comparing different doses Data were available to compare duloxetine at 40 mg (less than the licensed dose) with 80 mg. with more among those taking duloxetine. The full evidence profiles and associated forest plots can be found in Appendix 16c and Appendix 19c. respectively. There were no statistically or clinically important differences between the doses on either efficacy or acceptability and tolerability outcomes. although there were few trials. Those taking duloxetine also experienced a small average weight loss compared with those on placebo.37). However. One study comparing duloxetine at different doses included a continuation phase for those who achieved at least partial response to acute-phase treatment (defined as 365 . ELI LILLY HMAQ. respectively). 3. Evidence from the important outcomes and overall quality of evidence are presented in Table 78.77 to 1.

9 ( 2.78 (0.04* *The full data for these outcomes for different doses are shown in the forest plots Dul 01. 1663 Dul 02. Dul 01. 2895 K n 11.366 Nonremission Depression Leaving related pain treatment (average pain early in last 24 hours) WMD 0.11 Table 76: Summary evidence profile for duloxetine versus placebo (acute phase) Pharmacological interventions Mean depression change scores at endpoint Nonresponse Clinicianrated effect size WMD 1.24) (66 versus 51%) Moderate K n 10. 2249 K n 12.83 (0. 1763 Dul 02.02 2.01.22 (1.74 to 0.34 (0.03 and Dul 01.44 to 1. 2789 RR 1.38) Low K n 8.3 versus 11.07 Leaving treatment early due to side effects Leaving treatment early due to lack of efficacy N reporting side effects Weight change (kg) RR 0.02 (0.2%) Moderate K n Dul 01.09 Dul 02.91 to 1.09 WMD 0.74 ( 1.6 versus 67.79 to 0.5%) Moderate K n 6.95) (10 versus 5%) RR 0.83) (51. 2647 Dul 02.35) RR 0. 3078 Forest plot number Dul 01.22 to 0.04 Moderate Moderate RR 1.34) Moderate K n Dul 01.15) (26.02* Dul 01. .4%) RR 2. 2921 Dul 02.66 to 2.3%) Quality of evidence Moderate Low Number of studies/ participants K n 10. 583 K n 11.12 to 1.54) (7.69 ( 1 to 0.18 (1.06* 11.13 to 0.05 in Appendix 19c.9 versus 28.87) (62 versus 75.

Pharmacological interventions Table 77: Summary evidence profile for duloxetine versus placebo (continuation phase for partial responders) Mean depression change scores at endpoint 80 mg Clinician. n 152 120 mg WMD 0.54) (77 versus 87%) (4 versus 5%) Low Moderate 1. except venlafaxine which was more effective on mean change scores at endpoint (although the effect size was small and not quite statistically significant).78 to 1. respectively.04 RR 3.51 (0. Evidence from the important outcomes and overall quality of evidence are presented in Table 79. The quality of the evidence was low or very low.65) (5 versus 1%) Low 280 K 1.2 Clinician-rated ( 1.02 Dul 07.5 to 0.02) (0.75 to 1. This showed no difference between the doses.28 to 2.WMD 1 rated effect size ( 2.04 30% decrease in baseline HAMD scores) (PERAHIA2006B).01 Dul 07. Duloxetine was less acceptable to patients. Duloxetine versus other antidepressants Data were available to compare duloxetine with paroxetine.96 (0.03 Dul 07. fluoxetine. n 150 K Dul 07. There was no difference between duloxetine and other antidepressants.5) Quality of evidence Number of studies/ participants Forest plot number Low K 1.88 RR 0.81 to 1. n 140 K Dul 07.68) (1 versus 1%) Low 275 K 1.73) (82 versus 87%) (5 versus 5%) Low Moderate 1.38) effect size Quality of evidence Number of studies/ participants Forest plot number Low K 1. n 142 RR 0.94 RR 0.84 (0.34 to 2. n 152 K 2.08) (0.02 Dul 07. The full evidence profiles and associated forest plots can be found in Appendix 16c and Appendix 19c.01 Dul 07.4 to 30. and more people taking duloxetine left specifically because of adverse reactions. However. n 142 K 2. n Leaving treatment early Leaving treatment early due to side effects Leaving treatment early due to lack of efficacy RR 1 (0. n RR 0. as measured by the number leaving treatment early.03 Dul 07. escitalopram and venlafaxine. 367 .06 to 15.

91) (62 versus 65%) (57 versus 59%) (25 versus 30%) (5 versus 10%) Quality of evidence Very low Moderate Number of studies/ participants K 1.96 RR 0.31) Moderate 2. n Dul 04.35 ( 1 to 0.04 RR 0.45 to 1. n 647 K 1.84 to 1.47 (0.03 Dul 04.99 (0.99 (0. n 352 K 2. n 323 1.25 to 3. n 341 K Forest plot number Dul 03.73 RR 0.01 Dul 04.43 to 2.97 RR 0.03) RR 1.57 to 0.92 to 1.82 RR 0.07) (0. n 352 K 2.87) 1 versus 1% Low 1. n 325 Dul 03.11) (0.24) (0.58 ( 0. n 400 K .24 to 0.01 Dul 03.95) (0.9 to 1.07) (85 versus 86%) Moderate K 2. n 353 K 2.98 (0.62 to 1.43) (0.87 to 2.31) (62 versus 59%) (72 versus 63%) (35 versus 47%) (12 versus 15%) Quality of evidence Very low Low Number of studies/ participants K 2.89 to 1. n Table 78: Summary evidence profile for duloxetine comparing different doses (acute phase) Mean depression change scores at endpoint Non-response Pharmacological interventions 40 mg versus 80 mg Clinician-rated effect size WMD 0.19 ( 0.91 to 1. n 647 Dul 04.84 to 1.83 ( 0.3) Moderate 647 K 1.368 Non-remission Leaving treatment early RR 0.02 30 mg versus 60 mg Clinician-rated effect size WMD 0.05 RR 1.08) (0. n 352 Leaving treatment early due to side effects Leaving N reporting treatment side effects early due to lack of efficacy Weight change (kg) WMD 0.77 Not reported (0. n 647 Very low K 1.69 to 0. n 647 K Very low Low K 1.09) RR 0.1) (73 versus 74%) Low 647 K 1.02 – RR 0.05 WMD 0. n 352 – Moderate Low – Low K 2.15 RR 0.

04 Dul 04.69 to 0.5) (0. n Low Very low WMD 0.03 Dul 04.28 to 1.05 80 mg versus 120 mg Clinician-rated effect size Low 2.53) Dul 03.9 to 1.85 to 1.12 (0. n 384 K 2.44) (3 versus 2%) Very low 384 K 2.05 Pharmacological interventions 369 .83 to 1.23) (0.04 Dul 04.44 to 3.45 to 5.01 RR 1.03 Dul 04. n 384 K 2.15 RR 1.24) (35 versus 31%) (55 versus 55%) (12 versus 10%) (4 versus 4%) Quality of evidence Very low Low Number of studies/ participants Dul 03.68) RR 1.01 Dul 04.03) (0.02 Dul 04. n 384 K 2.01 Dul 03.02 Dul 03.Forest plot number RR 1.08 ( 0. n 384 K 2. n Low 384 Low K 1. n 381 K Forest plot number Dul 03. n 186 RR 1.03 Dul 04.13 RR 1.65 to 2.02 Dul 04.7 ( 0.01 Dul 04.03 K 2.4) (49 versus 41%) WMD 0.2 (0.01 Dul 03.56 (0.02 Dul 04.

152 Dul 08.26) (82 versus 77%) Low K n 1.29 (0.03 Dul 08. respectively. 152 Very low K n 1. Two studies comparing duloxetine with other antidepressants included a continuation phase for those who achieved at least partial response to acute-phase treatment (defined as 30% decrease in baseline HAMD scores) (DETKE2004. The full evidence profiles and associated forest plots can be found in Appendix 16c and Appendix 19c. on the number of people reporting side effects or on weight change. there was no difference between duloxetine and other antidepressants on numbers leaving treatment early because of lack of efficacy.58) (0. low or very low.Pharmacological interventions Table 79: Summary evidence profile for duloxetine comparing different doses (continuation phase for partial responders) Mean Leaving depression treatment change scores early at endpoint* 80 mg versus 120 mg Clinicianrated effect size Quality of evidence Number of studies/ participants Forest plot number WMD 0. The quality of the evidence was low. PERAHIA2006B). Both studies compared duloxetine with paroxetine. respectively.01 Dul 08. Evidence from the important outcomes and overall quality of evidence are presented in Table 80.18 to 0. Only one outcome was reported by both studies. Evidence from the important outcomes and overall quality of evidence are presented in Table 81.02 Dul 08.76 (0.03 to 2.91 to 1. This showed no difference between the doses. largely because of the selective population included in the studies. The quality of the evidence was moderate.04 *Change from end of acute phase.42) (3 versus 4%) Leaving treatment early due to lack of efficacy RR 0.49) (1 versus 5%) Very low K n 1. 152 Leaving treatment early due to side effects RR 0. The full evidence profiles and associated forest plots can be found in Appendix 16c and Appendix 19c. 370 . 150 Low K n 1.8 RR 1.07 ( 2.13 to 4.

26) (61 versus 60%) 1119 K 5.44 to 0.81 to 1. n Dul 05.94 to 1.74) Low 3145 Dul 05.89 to 1.02 Pharmacological interventions Non-remission RR 1.01 Dul 05.06 (0.66 ( 0. n Forest plot number Dul 05.9 to 1.72 to 1.27) (54 versus 51%) Continued 667 648 All Mean depression change scores at endpoint WMD 0.1 ( 3.99 (0.14 to 0.95 to 1.10) (56 versus 56%) Dul 05.01 (0. n 1184 K 2.02 RR 1.19 ( 0.Table 80: Summary evidence profile for duloxetine versus other antidepressants (acute phase) Paroxetine WMD 0.04 (0.94 to 1. n Dul 05.61 to 1.17) (49 versus 46%) Quality Low Number of studies.88 to 1.64) (40 versus 32%) Very low K 2.26) (44 versus 43%) Low 3208 Dul 05.83) WMD 0. n 222 Low RR 0.06 (0.61) (61 versus 52%) K 5.56 to 2.81) Quality Low Number of studies.02 RR 1.02 to 2.01 RR 1.02 (0.36) (53 versus 53%) Dul 05.99 (0. n Dul 05.02 RR 0.16) (59 versus 57%) Moderate K 3. participants K 12.2 ( 1.11) (58 versus 56%) 371 .03 to 0.92 to 1.06 ( 0.14) Moderate K 2. n 1200 K 2.93) Fluoxetine Escitalopram Venlafaxine WMD 1.01 RR 1. n 217 K 3.05 (0.02 RR 1. participants K 12.01 RR 1.21 (0.23 (0. n Forest plot number Dul 05.01 Non-response RR 1. n 1096 Very low Low WMD 1.

01 to 1.42) (15 versus 9%) Moderate K 2.01 RR 2. n 222 Very low Low K 2.01 to 1.72) (35 versus 26%) Moderate K 2. participants K 11.2 to 1.32 (0.21 (1.03 Dul 05.78) (32 versus 21%) K 5.03 Leaving treatment early RR 1.62 (0.03 RR 1.01 RR 1.97 to 2. participants K 10.01 RR 3.01 RR 1.45) (32 versus 24%) Quality Low Number of studies. n Dul 06.27) (32 versus 37%) Dul 05. participants K 12.03 RR 1. n 1200 K 2. n Dul 05.93) (9 versus 7%) Low 2795 K 5. n Dul 06.04 to 2.01 Leaving treatment early due to side effects RR 1. n 222 K 3.03 RR 1. n Forest plot number Dul 05.45) (29 versus 24%) Moderate 2914 Dul 06.3 (0. n 1119 Very low Low Low K 2.67 to 10.87 (0.3) (9 versus 4%) Very low K 2.64 (0.58 (1. n 103 K 5.37 (1. n 667 667 667 Fluoxetine Escitalopram Venlafaxine All Quality Low Pharmacological interventions Number of studies.09 to 1.372 Table 80: (Continued) Paroxetine Moderate 3208 Dul 05.54 (1. n 1200 K 2. n .42 to 25. n 1200 Dul 06.74) (10 versus 3%) Very low K 1.27 (1.59 to 1. n 825 825 RR 0. n Forest plot number Dul 06.99) (10 versus 7%) Quality Moderate Number of studies.9 to 1.

04 WMD 0.Forest plot number RR 2.09 to 0.70 to 1.01 ( 0. n 825 K – Very low Very low 2.6 to 8. n Dul 06.99 to 1. n Pharmacological interventions Forest plot number Dul 06.04 WMD 0.39 ( 0.68) (3 versus 3%) Quality Moderate Number of studies. participants K 9.08 to 1.51 to 1.03 RR 1.06 ( 1. n Dul 06.04 WMD 0 ( 0.72) K 5. n 834 Low K 1.03) Moderate 2207 Dul 06.07 (0.15) (71 versus 65%) Moderate 2517 Dul 06.97 (0. n Dul 06.78) (2 versus 1%) Very low 2341 Dul 06.06 579 667 667 Not reported RR 0.02 Dul 06.03 RR 1.96 to 1.06 547 547 RR 0. n 849 – K 2. n Dul 06.86) Moderate K 2.02 (0.02 (0.04 Weight change (kg) WMD 0 ( 0.2) Low K 1.03 to 0.29 (0. n Dul 06.98 to 1. n Forest plot number Dul 06. n 1200 K 1.03) Quality Moderate Number of studies.06 98 Dul 06. n 103 Low Low K 1.06 K 4.53) (5 versus 6%) RR 1.09) (88 versus 87%) K 3.09 (0.02 Leaving treatment early due to lack of efficacy RR 1.12) (89 versus 91%) – Dul 06.88 (0.07) (79 versus 76%) Quality Moderate Number of studies.03 RR 0.85 to 1.88 to 1.02 Dul 06.74 to 0. n Dul 06.11) (86 versus 87%) Low K 2.99 (0.24 (0.03 to 0.05 373 .02 Dul 06.95) (3 versus 3%) Dul 06.52 to 2. participants K 7.03 N reporting side effects RR 1.02 Dul 06. n Forest plot number Dul 06.04 WMD 0. participants K 8.

participants Forest plot number Leaving treatment early due to side effects Quality Number of studies. One study comparing duloxetine with other antidepressants included a continuation phase for all those entering the study regardless of response during the acute phase of the study (WADE2007). participants Forest plot number Leaving treatment early Quality Number of studies. Evidence from the important outcomes and overall quality of evidence are presented in Table 82.04 from end of acute phase.Pharmacological interventions Table 81: Summary evidence profile for duloxetine versus other antidepressants (continuation phase for partial responders) Paroxetine Mean depression change scores at endpoint* Quality Number of studies. which was not clinically important.81 to 1. n 141 Dul 09.6) (5 versus 1%) Very low K 2.01 RR 0. n 140 Dul 09.03 RR 0.31) (1 versus 3%) Very low K 1.02 RR 2. There was a small difference in favour of escitalopram in efficacy measures. respectively.66) Low K 1.08) (82 versus 87%) Low K 1. 374 . participants Forest plot number *Change WMD 0. n 141 Dul 09.94 (0.05 to 5. participants Forest plot number Leaving treatment early due to lack of efficacy Quality Number of studies. and the number of patients leaving treatment early specifically because of side effects favoured escitalopram.3 ( 1.7 to 11. This compared duloxetine with escitalopram.49 (0.06 to 1. n 286 Dul 09.84 (0. The full evidence profiles and associated forest plots can be found in Appendix 16c and Appendix 19c.

74 to 1.06 375 . n 294 Dul 11.54) (17 versus 9%) Low K 1.25 to 2.89 (1.65) (33 versus 28%) Very low K 1.01 RR 1.82 to 1.01 to 3.28) (1 versus 5%) Very low K 1.Pharmacological interventions Table 82: Summary evidence profile for duloxetine versus other antidepressants (continuation phase for all) Escitalopram Mean depression change scores at endpoint* Quality Number of studies. participants Forest plot number Leaving treatment early due to side effects Quality Number of studies.32 (0.06 to 1. n 294 Dul 11.16 (0.27 (0. participants Forest plot number *Change from end of acute phase. n 294 Dul 11. participants Forest plot number Leaving treatment early due to lack of efficacy Quality Number of studies. participants Forest plot number Non-remission Quality Number of studies.03 RR 1.02) (26 versus 20%) Low K 1.02 RR 1.34 ( 0. participants Forest plot number Non-response Quality Number of studies.13 (0. n 294 Dul 11. n 287 Dul 11.86 to 2. WMD 1. participants Forest plot number Leaving treatment early Quality Number of studies. n 294 Dul 11.72) (25 versus 22%) Very low K 1.04 RR 1.05 RR 0.93) Low K 1.

All included studies were published between 1990 and 2003 and were between 5 and 24 weeks’ long (mode 6 weeks). Overall the quality of the evidence was downgraded because of the highly selective patient populations in the trials. Information about each study along with an assessment of methodological quality is in Appendix 17c.491 trial participants and tolerability data from up to 2. Bruijn1996. Schatzberg2002. There appears to be no advantage for doses of duloxetine above the licensed dose of 60 mg. which also contains a list of excluded studies with reasons for exclusions. although there are few trials comparing higher doses. These provided efficacy data from up to 2. three-quarters of a point difference between the groups). It can cause weight gain and sedation.3 Mirtazapine Introduction **Mirtazapine is a noradrenaline and specific serotonin antidepressant (NaSSA) that blocks presynaptic alpha 2 receptors on both NA and 5HT neurones and also blocks postsynaptic 5HT2 (less sexual dysfunction but possible worsening of the symptoms of obsessive-compulsive disorder) and 5HT3 (less nausea) receptors. Nine studies compared mirtazapine with TCAs and related antidepressants (Bremner1995. Zivkov1995).10.34] that is. were excluded because more than 50% of participants left treatment early) and 11 were excluded by the GDG. Leinone1999. There was no advantage found for increasing the dose for partial responders.637 participants. Since duloxetine is still in patent its acquisition costs are relatively high compared with antidepressants available in generic form (see Section 10. five compared it with SSRIs (Benkert2000. 125Study IDs in title case refer to studies included in the previous guideline. The difference in endpoint depression scores compared with placebo is small. VanMoffaert1995. References for these studies are in Appendix 18. Wade2003. Marttila1995. one was from primary care and in the other three 124Details of standard search strings used in all searches are in Appendix 8.Pharmacological interventions Clinical summary There does not seem to be any advantage for duloxetine over other antidepressants.125 Twenty-five studies were found in a search of electronic databases and details of a study in press were provided by Organon Laboratories Ltd (Wade2003). with evidence for some outcome-comparison combinations being downgraded further largely because of low numbers of trials. Smith1990. In five studies participants were described as inpatients.8.13 to 0. and there does not seem to be an important reduction in pain associated with depression in those trials that reported this measure (WMD 0. Studies considered124.2). and no trials comparing 60 mg with higher doses. in six as outpatients. 10. Halikas1995. Wheatley1998). 376 . Wade2003. Fifteen studies were included (although the efficacy data from one of these. Richou1995. Mullin1996. and one with venlafaxine (Guelfi2001).74 [ 1.

N 854.23 to 0. 95% CI. N 2440.03. clomipramine.1.05. 95% CI. Clinical evidence statements126 Effect of treatment on efficacy outcomes There is no difference between the efficacy of mirtazapine and other antidepressants for which comparisons were available: There is evidence suggesting that there is no clinically important difference between mirtazapine and other antidepressants on: ● increasing the likelihood of achieving a 50% reduction in symptoms of depression by the end of treatment as measured by the HRSD (K 14127.314.05). N 819. Random effects SMD 0. There is evidence suggesting that there is a statistically significant difference favouring mirtazapine over other antidepressants on increasing the likelihood of achieving remission by the end of treatment as measured by the HRSD. all participants were 65 years of age or older. imipramine. N 2. –0. fluoxetine.92. 95% CI.27 to 0. –0.91. 95% CI. paroxetine and venlafaxine. SMD –0.93.05) ● reducing symptoms of depression in outpatients by the end of treatment as measured by the HRSD or the MADRS (K 6. Data were available to compare mirtazapine with amitriptyline.Pharmacological interventions it was either not clear from where participants were sourced or they were from mixed sources. In outpatients there is evidence suggesting that there is a statistically significant difference favouring mirtazapine over other antidepressants on increasing the 126The 127One forest plots can be found in Appendix 19c. Similar results were found in sub-analyses by antidepressant class. 0. –0. 95% CI. N 387. SMD –0. but the size of this difference is unlikely to be of clinical importance (K 4. doxepin.24) ● increasing the likelihood of achieving remission in outpatients by the end of treatment (K 2. 0. 95% CI. In one study (Schatzberg2002). but the size of this difference is unlikely to be of clinical importance (K 4. Effect of setting on efficacy outcomes There is evidence suggesting that there is no clinically important difference between mirtazapine and other antidepressants on: ● reducing symptoms of depression by the end of treatment in inpatients as measured by the HRSD or MADRS (K 5. SMD –0. 95% CI.84 to 1.11 to 0. 0. other than for SSRIs: There is evidence suggesting that there is a statistically significant difference favouring mirtazapine over SSRIs on reducing symptoms of depression by the end of treatment.03). RR 0. N 888.00). trazodone.01) ● reducing symptoms of depression by the end of treatment as measured by the HRSD or the MADRS (K 14.15 to 0. RR 0.81 to 1. N 915. citalopram.13.99).83 to 0. –0. 377 .2 mg of mirtazapine. study (Wade2003) was removed because 50% of participants left the study early. RR 0. Studies reported mean doses of between 22 and 76.

There is strong evidence suggesting that there is a clinically important difference favouring mirtazapine over other antidepressants on reducing the likelihood of patients leaving treatment early due to side effects (K 15. RR 0. Acceptability and tolerability of treatment Mirtazapine appears to be as acceptable to patients as other antidepressants. It is not licensed for use in older adults.93 to 1. mirtazapine has a statistical advantage over SSRIs in terms of reducing symptoms of depression. 2003). N 2637.8. 95% CI. No data were available to determine efficacy in patients in primary care. There is evidence suggesting that there is no clinically important difference between mirtazapine and other antidepressants on reducing the likelihood of patients reporting side effects (K 6. N 957. sweating.Pharmacological interventions likelihood of achieving a 50% reduction in symptoms of depression by the end of treatment as measured by the HRSD.88. It may also lower serum potassium (The Association of the British Pharmaceutical Industry.55 to 0. 0. Findings were similar in sub-analyses by setting and class of antidepressant. except that fewer patients leave treatment early due to side effects: There is evidence suggesting that there is no clinically important difference between mirtazapine and other antidepressants on reducing the likelihood of leaving treatment early for any reason (K 15.78 to 1).69.86. although mirtazapine is as effective as other antidepressants.** 10. RR 0. 0.4 Reboxetine Introduction **Reboxetine is a relatively selective.99. dizziness.73 to 1). Clinical summary There is no difference between mirtazapine and other antidepressants on any efficacy measure. but the size of this difference is unlikely to be of clinical importance (K 6. 95% CI.87). However. 95% CI. In inpatients there is insufficient evidence to determine whether there is a clinically important difference between mirtazapine and other antidepressants on increasing the likelihood of achieving a 50% reduction in symptoms of depression or on achieving remission. noradrenergic reuptake inhibitor. it may have an advantage in terms of reducing side effects likely to lead to patients leaving treatment early.05). Therefore. RR 0. 378 . 95% CI. 0. dry mouth and constipation (Holm & Spencer. but the difference is not clinically important. In addition. 1999). Side effects include insomnia. RR 0. although in terms of achieving remission mirtazapine appears to have a statistical though not clinical advantage. although this is not the case for patients reporting side effects or leaving treatment early for any reason. N 1253. 0. there is strong evidence that patients taking mirtazapine are less likely to leave treatment early because of side effects. N 2637.

59 to 0. All included studies were published between 1997 and 2002 and were between 4 and 8 weeks’ long (mean 6. Clinical evidence statements for reboxetine compared with other antidepressants131 Effect of treatment on efficacy outcomes There is evidence suggesting that there is no clinically important difference between reboxetine and other antidepressants on: ● increasing the likelihood of achieving a 50% reduction in symptoms of depression as measured by the HRSD (K 5. 95% CI. Versiani2000B) being included and two excluded. Massana1999). fluoxetine and placebo. where participants received a dose of 6 mg. Katona1999. 131The forest plots can be found in Appendix 19c. Data were available to compare reboxetine with desipramine. RR 0. Berzewski1997.66 weeks).Pharmacological interventions Studies considered128. Massana1999. 0. Berzewski1997. There is some evidence suggesting that there is a clinically important difference favouring reboxetine over placebo on increasing the likelihood of achieving remission by the end of treatment (K 1. N 254. Katona1999) and two with SSRIs (Andreoli2002. all participants were aged 65 years and over. Clinical evidence statements for reboxetine compared with placebo130 Effect of treatment on efficacy outcomes There is strong evidence suggesting that there is a clinically important difference favouring reboxetine over placebo on increasing the likelihood of achieving a 50% reduction in symptoms of depression as measured by the HRSD (K 3. 95% CI. imipramine. 130The forest plots can be found in Appendix 19c. Acceptability and tolerability of treatment There is insufficient evidence to determine whether there is a clinically important difference between reboxetine and placebo on any measure of acceptability or tolerability.01) 128Details of standard search strings used in all searches are in Appendix 8. three with TCAs (Ban1998. which also contains a list of excluded studies with reasons for exclusions. doses were between 8 and 10 mg of reboxetine.76 to 1.73). 379 . 0.129 Eight studies were found in a search of electronic databases. In one (Katona1999).51 to 0. with six (Andreoli2002. 0. Ban1998. Ban1998. Three studies compare reboxetine with placebo (Andreoli2002. Information about each study along with an assessment of methodological quality is in Appendix 17c.87. N 479. N 1068. Apart from Katona1999. These provided efficacy and tolerability data from up to 1. RR 0.87).068 trial participants. References for these studies are in Appendix 18. 95% CI.61. Versiani2000B). RR 0.71. In two studies participants were described as inpatients and in the other three it was either not clear from where participants were sourced or they were from mixed sources. 129Study IDs in title case refer to studies included in the previous guideline.

is associated with a high incidence of discontinuation symptoms (see Section 11. Information about each study along with an assessment of methodological quality is in Appendix 17c. 0. Fifteen additional studies were identified from new searches and four from another review (Einarson et al.8) and is more toxic than the SSRIs in overdose (see Section 11.09) reducing symptoms of depression by the end of treatment as measured by the HRSD or MADRS (K 3. None of these studies met the inclusion criteria set by the GDG.24 to 0. Clinical summary Reboxetine is superior to placebo and as effective as other antidepressants in the treatment of depression. 0. 95% CI. five of which 132Details of standard search strings used in all searches are in Appendix 8. 95% CI.. SMD –0. both of which met inclusion criteria.. 10. 380 . which also contains a list of excluded studies with reasons for exclusions. The Smith and colleagues’ (2002) review included 31 studies of which nine did not meet the inclusion criteria set by the GDG.5 Venlafaxine Introduction **Venlafaxine was the first of the new generation dual-action antidepressants. Venlafaxine has a broad range of side effects similar to those of TCAs and SSRIs. –0. N 895. 2002) as the basis of this review. It inhibits the reuptake of both serotonin and noradrenaline in the same way as TCAs.9 to 1. RR 0. At the standard dose of 75 mg it is an SSRI. References for these studies are in Appendix 18. and details of ten additional unpublished studies were provided by Wyeth Laboratories.98. 133Study IDs in title case refer to studies included in the previous guideline.07). There is insufficient evidence to comment on reboxetine’s tolerability compared with placebo or alternative antidepressants.9). 95% CI. with dual action emerging at doses of 150 mg and above.133 The GDG used an existing review (Smith et al. There is insufficient evidence to determine whether there is a clinically important difference between reboxetine and other antidepressants on reducing the likelihood of leaving treatment early for any reason or on reducing the likelihood of leaving treatment early due to side effects. At higher doses it also inhibits dopamine reuptake.Pharmacological interventions ● ● increasing the likelihood of achieving remission by the end of treatment (K 4. It can increase blood pressure at higher doses.96.84 to 1.09.8. n 895. 1999).06). N 618. Studies considered132. Acceptability and tolerability of treatment There is evidence suggesting that there is no clinically important difference between reboxetine and other antidepressants on increasing the likelihood of patients reporting side effects (K 4. Two studies were sourced from other reviews in this chapter. RR 0.

Data were available to compare venlafaxine with clomipramine. imipramine. Doses ranged from 75 mg to 375 mg.Pharmacological interventions met inclusion criteria. Bielski2003. 349Wyeth. Smeraldi1998. trazodone. 671Lenox-Smith.6 (various HRSD versions). studies with a dose of 75 mg were analysed separately (102Tsai. In the remaining six. 428Casabona. Results are presented only where clinically important differences were found. Eight studies (102Tsai. 671Lenox-Smith. Random effects RR 0. 626Kornaat. McPartlin1998. Samuelian1998. Smeraldi1998. Guelfi2001. 332Rickels. Tylee1997. Smerladi1998) participants were aged 64 years and over. Schweizer1994. citalopram. Bielski2003. Rudolph1999. Some participants in one study (Guelfi2001) received the comparator treatment (mirtazapine) at a dose higher than BNF limits. Rudolph1999. N 4198. Poirier1999. In three (Mahapatra1997. Mahapatra1997.92. Mean HRSD scores at baseline ranged from 22. Costa1998. Hackett1996. Montgomery2002.02) 134The forest plots can be found in Appendix 19c. Bielski2003.03 weeks). escitalopram. Where this gave heterogeneity. Tylee1997). Benkert1996. Together. fluoxetine. Lecrubier1997. Clinical evidence statements134 Effect of treatment on efficacy Venlafaxine is no more effective in treating depression than other antidepressants: There is evidence suggesting that there is no clinically important difference between venlafaxine and other antidepressants on: ● increasing the likelihood of achieving a 50% reduction in symptoms of depression as measured by the HRSD (K 23. Three studies were of inpatients. All included studies were published between 1994 and 2003 and were between 4 and 13 weeks’ long (mean 8. these provide tolerability data from up to 5. 015Schatzberg. Montgomery2002. 102Tsai. McPartlin1998. 16 of outpatients and four were undertaken in primary care. Silverstone1999.4 to 30. Studies reported mean doses equivalent to at least 100 mg of amitriptyline. Montgomery2002. Tzanakaki2000). paroxetine and mirtazapine. Mahapatra1997. 428Casabona. Costa1998. Alves1999. Hackett1996. 428Casabona. sub-analyses were performed removing this study. Clerc1994. Dierick1996. In addition. 381 . Lecrubier1997. Dierick1996. 349Wyeth.063 participants and efficacy data from up to 4. A sub-analysis was performed by dose of venlafaxine. 428Casabona. Thus a total of 33 studies were excluded from this review with 29 trials being included (014Nemeroff.198 participants. dothiepin/ dosulepin. with studies achieving a maximum dose of no more than 150 mg classified as low dose (102Tsai. 95% CI. Alves1999. Benkert1996. Samuelian1998. 015Schatzberg. Poirier99. Tylee1997) and those achieving a minimum dose of no less than 150 mg classified as high dose (332Rickels. it was either not clear from where participants were sourced or they were from mixed sources. Tzanakaki2000). McPartlin98. Hackett1996. Silverstone1999) used ‘extended release’ (XR) venlafaxine and the remainder ‘immediate release’ (IR) venlafaxine. 0. Cunnigham1994. Clerc1994. Guelfi2001.83 to 1.

005) ● remission (K 19. SMD –0.53) or on increasing the likelihood of achieving remission (K 2. Random effects SMD –0.002). RR 0. 95% CI. RR 0. –0. 95% CI. 95% CI. but the size of this difference is unlikely to be of clinical importance (K 20.58. 0. 95% CI. –0. 0. compared with SSRIs. 0. N 773.02). 0.10. RR 0. –0. N 392. SMD –0. Random effects RR 0. SMD –0.62). N 3692. N 3637.84 to 1. fluoxetine and mirtazapine. Inpatients: There is evidence suggesting that there is no clinically important difference between venlafaxine and other antidepressants on reducing symptoms of depression in inpatients by the end of treatment as measured by the HRSD or MADRS (K 3.96.9 to 1. N 2741. 0. data were available to compare it with imipramine.01). N 744. N 3268. 95% CI. venlafaxine is more effective in inpatients: There is some evidence suggesting that there is a clinically important difference favouring venlafaxine over SSRIs on: ● reducing symptoms of depression in inpatients by the end of treatment as measured by the HRSD or MADRS (K 1. There is evidence suggesting that there is no clinically important difference between venlafaxine and TCAs on reducing symptoms of depression by the end of treatment as measured by the HRSD or MADRS (K 6. There is insufficient evidence to determine if there is a clinically important difference between venlafaxine and TCAs on increasing the likelihood of patients achieving a 50% reduction in symptoms of depression as measured by the HRSD or MADRS (K 6. –0. However.38).Pharmacological interventions ● increasing the likelihood of achieving remission as measured by the HRSD (K 20.46 to 0.02).09) 382 . There is evidence suggesting that there is a statistically significant difference favouring venlafaxine over SSRIs on reducing symptoms of depression by the end of treatment but the size of this difference is unlikely to be of clinical importance (K 13. N 225. 95% CI. N 383.07 to –0. Random effects RR 1. SMD –0.45 to 1.91 to 1.12.95.04.02).71 to 1.85. There is evidence suggesting that there is a statistically significant difference favouring venlafaxine over other antidepressants on reducing symptoms of depression. 95% CI. 95% CI. 95% CI.17). There is insufficient evidence to determine whether there is a clinically important difference between venlafaxine and other antidepressants on either increasing the likelihood of achieving a 50% reduction in symptoms of depression (K 3. 0. Similar results were found in sub-analyses by class of antidepressant: There is evidence to suggest that there is no clinically important difference between venlafaxine and SSRIs on increasing the likelihood of achieving: ● a 50% reduction in symptoms of depression (K 16. N 3849. –1. Effect of setting on treatment efficacy To assess the efficacy of venlafaxine in inpatients.04. N 67. 95% CI.09. Random effects RR 0.27 to 0. 95% CI.71 to 1.17 to –0.92.91.15 to –0.

95% CI.05).83.25 to –0. SMD –0.88 to 1. 95% CI. when venlafaxine was compared with SSRIs there is evidence suggesting that there is no clinically important difference between venlafaxine and SSRIs on increasing the likelihood of achieving remission in outpatients (K 12. paroxetine and fluoxetine in primary care. RR 0.06). –0. Outpatients: Data from studies of venlafaxine in outpatients were available to make comparisons with imipramine. SMD –0. In outpatients. 95% CI. There is insufficient evidence to determine if there is a clinically important difference between venlafaxine (75 mg) and SSRIs on increasing the likelihood of patients 383 . SMD –0. RR 0. but the size of these differences is unlikely to be of clinical importance on: ● increasing the likelihood of achieving a 50% reduction in symptoms of depression by the end of treatment (K 9. There is evidence suggesting that there is no clinically important difference between venlafaxine and other antidepressants on reducing symptoms of depression by the end of treatment as measured by the HRSD or MADRS (K 3. 0. 0.75 to 0.74 to 0. N 68. There is some evidence suggesting that there is a clinically important difference favouring venlafaxine over other antidepressants on increasing the likelihood of achieving a 50% reduction in symptoms of depression in outpatients as measured by the HRSD (K 11. 95% CI. –0. 95% CI. 95% CI. fluoxetine and paroxetine. 95% CI. 0.93).21 to 0.26 to –0. There is evidence suggesting that there is a statistically significant difference favouring venlafaxine over other antidepressants on reducing symptoms of depression in outpatients by the end of treatment as measured by the HRSD or MADRS.17. However. but the size of this difference is unlikely to be of clinical importance (K 9. there is evidence suggesting that there are statistically significant differences favouring venlafaxine over SSRIs on the following outcomes.15.07. 0. 95% CI.89 to 1.08). N 1572. Results were similar against TCAs alone. 0.11). Effect of dose on treatment efficacy Venlafaxine at 75 mg: Data were available to compare venlafaxine at 75 mg with fluoxetine and paroxetine.02).39 to 0. RR 0. –0. Primary care: Data were available to compare venlafaxine against imipramine. N 1804. N 824.Pharmacological interventions ● increasing the likelihood of achieving remission in inpatients as measured by the HRSD (K 1.92).95.96) ● reducing symptoms of depression in outpatients by the end of treatment (K 7. N 2023. clomipramine. N 1775.85.98. N 995. RR 0. There is evidence suggesting that there is no clinically important difference between venlafaxine and SSRIs on increasing the likelihood of achieving remission (K 3. RR 0.60. N 2199.

98.9 to 1. RR 0.09) ● reducing symptoms of depression as measured by the HRSD at the end of treatment (K 3.26). 0.12).87. There is evidence suggesting that there is no clinically important difference between venlafaxine ( 150 mg) and other antidepressants on increasing the likelihood of achieving remission (K 9.01. Random effects RR 0. 95% CI. Low-dose venlafaxine (mean 150 mg): There is insufficient evidence to determine if there is a clinically important difference between venlafaxine ( 150 mg) and other antidepressants on increasing the likelihood of patients achieving a 50% reduction in symptoms of depression as measured by the HRSD or MADRS (K 12. Random effects RR 1. 0. N 822.94. 95% CI.78 to 1.18 to 0.21 to 0.19 to –0. 95% CI.97 to 1. 384 .05). N 792. 95% CI.88 to 1. 95% CI. N 2125. RR 0. 0. N 2418. 0. –0. Random effects SMD 0. 0.6 to 1. 95% CI.86. SMD –0. 95% CI. There is evidence to suggest that there is no clinically important difference between venlafaxine (75 mg) and SSRIs on: ● increasing the likelihood of patients achieving remission as measured by the HRSD or MADRS (K 4. 0. Acceptability and tolerability of treatment There is evidence suggesting that there is no clinically important difference between venlafaxine and other antidepressants on: ● reducing the likelihood of leaving treatment early for any reason (K 23. N 882.79 to 1. There is evidence suggesting that there is a statistically significant difference favouring venlafaxine ( 150 mg) over other antidepressants on reducing symptoms of depression as measured by the HRSD or MADRS at the end of treatment but the size of this difference is unlikely to be of clinical importance (K 11. Random effects RR 0. Results were similar in sub-analyses by antidepressant class. Random effects RR 0.88 to 1.23) ● increasing the likelihood of achieving remission (K 6.08. –0. Results were similar in sub-analyses by antidepressant class.06).11.06).98. 95% CI.08) ● reducing the likelihood of patients reporting adverse events (K 21. N 2256. N 4196. 95% CI. –0. N 807. 95% CI. N 3757.03. 0. N 882. There is evidence suggesting that there is no clinically important difference between venlafaxine ( 150 mg) and other antidepressants on: ● reducing symptoms of depression (K 6.Pharmacological interventions achieving a 50% reduction in symptoms of depression as measured by the HRSD or MADRS (K 4.28). N 706.03). RR 1.72 to 1. 0. SMD –0. 95% CI. High-dose venlafaxine (mean 150 mg): There is insufficient evidence to determine if there is a clinically important difference between venlafaxine ( 150 mg) and other antidepressants on increasing the likelihood of patients achieving a 50% reduction in symptoms of depression as measured by the HRSD or MADRS (K 6. RR 0.98.02).

1.41). 1 to 1. N 1736. When compared with SSRIs: There is some evidence suggesting that there is a clinically important difference favouring SSRIs over venlafaxine on reducing the likelihood of outpatients leaving treatment early due to side effects (K 11. RR 0. the result against TCAs and SSRIs only were considered: There is some evidence suggesting that there is a clinically important difference favouring venlafaxine over TCAs and SSRIs on reducing the likelihood of inpatients leaving treatment early (K 2.48.04 to 1.94.92).16). RR 0.09). Primary care: There is evidence suggesting that there is no clinically important difference between venlafaxine and other antidepressants on: ● reducing the likelihood of leaving treatment early for any reason (K 4. 0.77 to 1. N 787.82 to 1. except for fluoxetine: There is evidence suggesting that there is a statistically significant difference favouring fluoxetine over venlafaxine on reducing the likelihood of patients reporting side effects.1) ● reducing the likelihood of outpatients reporting side effects (K 10. Outpatients: There is evidence suggesting that there is no clinically important difference between venlafaxine and other antidepressants on: ● reducing the likelihood of outpatients leaving treatment early for any reason (K 11. N 2.11).06. 1.95. N 2085. 0. Therefore.9995 to 1. but the size of this difference is unlikely to be of clinical importance (K 10. 95% CI. 0. RR 1. 95% CI.98 to 1. 95% CI. Acceptability and tolerability of treatment by setting Inpatients: To assess the efficacy of venlafaxine in inpatients. 95% CI. N 5063.61. N 235.Pharmacological interventions There is some evidence suggesting that there is a clinically important difference favouring other antidepressants over venlafaxine on reducing the likelihood of patients leaving treatment early due to side effects (K 27.16 to 1. This was because in the study comparing venlafaxine with mirtazapine. 0. Heterogeneity was a problem in the meta-analysis assessing the tolerability of venlafaxine against all antidepressants in inpatients. 95% CI. results were similar for venlafaxine compared with SSRIs.03.41 to 0. 385 . N 1148.15) ● reducing the likelihood of patients reporting adverse events (K 3.021. data were available to compare it with imipramine. In sub-analyses by antidepressant class. RR 1. fewer participants taking mirtazapine left the study early compared with those taking venlafaxine.21. 95% CI. RR 1. 0.08. RR 1. whereas this was not the case in other studies. N 1871. 95% CI. RR 1. RR 0. fluoxetine and mirtazapine.90). 95% CI.

95% CI. 1. 95% CI. N 674. RR 1.68 to 1. High-dose venlafaxine ( 150 mg): There is insufficient evidence to determine whether there is a clinically important difference between high-dose venlafaxine and other antidepressants on reducing the likelihood of leaving treatment early (K 6. There is evidence suggesting that there is no clinically important difference between high-dose venlafaxine and other antidepressants on reducing the likelihood of patients reporting side effects (K 6. N 768. Random effects RR 1. N 873. Random effects RR 1.06. N 2224.002 to 1.04.71 to 3. 95% CI. Clinical summary There are no clinically important differences between venlafaxine (at any dose) and other antidepressants on any efficacy outcome.001 to 1. 95% CI. 0.55). Low-dose venlafaxine ( 150 mg): There is evidence suggesting that there is no clinically important difference between low-dose venlafaxine and other antidepressants on reducing the likelihood of leaving treatment early (K 12.24). 0.19). N 2471. N 521.12. There is evidence suggesting that there is a statistically significant difference favouring other antidepressants over low-dose venlafaxine on reducing the likelihood of patients reporting side effects but the size of this difference is unlikely to be of clinical importance (K 12. 0. 95% CI. 0. 0. 1. RR 0. 95% CI. there is some evidence that patients 386 . N 2471.85 to 1.07.25).7 to 1. N 768. In sub-analyses by class of antidepressant. N 1190.61. 0. There is insufficient evidence to determine whether there is a clinically important difference between low-dose venlafaxine and TCAs on reducing the likelihood of leaving treatment early due to side effects. 95% CI. This was also the case for most acceptability and tolerability outcomes. 1.93. RR 1.12).05).95. 95% CI.75 to 1.16) ● reducing the likelihood of patients leaving treatment early due to side effects (K 3.41) or on reducing the likelihood of leaving treatment early due to side effects (K 7. RR 1.996 to 1. N 822. 0. RR 1. Random effects RR 1.05).25.7) ● reducing the likelihood of patients reporting side effects (K 3. RR 0. However. There is some evidence suggesting that there is a clinically important difference favouring other antidepressants over venlafaxine ( 150 mg) on reducing the likelihood of patients leaving treatment early due to side effects (K 12.48. 95% CI.91 to 1. results were similar except that: There is strong evidence that there is a clinically important difference favouring fluoxetine over low-dose venlafaxine on reducing the likelihood of leaving treatment early due to side effects (K 5.Pharmacological interventions Acceptability and tolerability of treatment by dose Venlafaxine at 75 mg: There is insufficient evidence to determine if there is a clinically important difference between venlafaxine (75 mg) and SSRIs on: ● reducing the likelihood of patients leaving treatment early (K 3. RR 1.15 to 2. 95% CI.

one small study of inpatients found that venlafaxine was superior to SSRIs on efficacy.1 Introduction **St John’s wort. Kalb2001. 136Study IDs in title case refer to studies included in the previous guideline. In addition. Volz2000. not be therapeutically equivalent. Many different branded preparations are available.9. has been used for centuries for medicinal purposes including the treatment of depression. 387 . Shelton2001. 2004). there was some evidence for increased efficacy compared with other antidepressants. an extract of the plant Hypericum perforatum. Philipp99. Laakmann98. Witte1995). four studies for a comparison with TCAs (Bergmann93. Information about each study along with an assessment of methodological quality is in Appendix 17c. Wheatley97. 2004). Most commercial products are standardised with respect to hypericin content. It is not licensed as a medicine in the UK but can be bought ‘over the counter’ from health food shops. but it is not known if this is the only active component.136 Forty studies were found in a search of electronic databases. Many clinically important drug interactions have been reported (Committee on Safety of Medicines. herbalists and community pharmacies. particularly when low-dose ( 150 mg) venlafaxine is compared with fluoxetine. one for a comparison with 135Details of standard search strings used in all searches are in Appendix 8. 10. xanthons and biflavonoids (Wagner & Bladt. Woelk2000). other than for inpatients. Ten studies were available for a comparison with placebo (Davidson02.** 10. but only on response. but its exact mode of action is unknown. the quantity and proportions of each constituent varies among batches (Wang et al. Schrader98. Hansgen1996. St John’s wort is known to contain at least ten constituents or groups of components that may contribute to its pharmacological effects (Linde & Mulrow. Lecrubier02. 1994). flavonoids.2 Studies considered135.9 ST JOHN’S WORT The following sections on St John’s wort marked by asterisks (**_**) are from the previous guideline and have not been updated except for style and minor clarification. with those taking venlafaxine being less likely to stop treatment early compared with TCAs and SSRIs. In common with all herbal preparations. St John’s wort may also cause photosensitivity. Philipp99.Pharmacological interventions taking venlafaxine are more likely to leave treatment early due to side effects. therefore. Individual brands or batches of the same brand may.9.. St John’s wort is licensed in Germany for the treatment of depression. with 19 being included and 21 being excluded by the GDG. Results were similar in sub-analyses by setting. In outpatients. These include naphthodianthrons. which also contains a list of excluded studies with reasons for exclusions. 10. References for these studies are in Appendix 18. 2000).

28% hypericin ● 4 125 mg (500 mg) Neuroplant ● 200–240 mg Psychotonin forte ● 3 30 drops Psychotonin (500 mg) ● 3 30 drops Hyperforat: 0. In addition.9.2 to 0. All included studies were published between 1993 and 2002 and were between 4 and 12 weeks’ long (mean 6. 388 . and six studies for a comparison with SSRIs (Behnke2002.450 to 0. RR 0.5 mg hypericin ● 3 to 6 300 mg (900 mg to 1800 mg) at 0.79. In one study (Harrer99). six studies with low doses of standard antidepressants were also included. Data from up to 1520 participants were available from studies comparing St John’s wort with placebo. Brenner00.12 to 0. In 16 studies participants were described as outpatients and in the other three it was either not clear from where participants were sourced or they were from mixed sources.12 to 0. 139Three studies (Davidson02.2% hypericin ● 900 mg to 1500 mg LI 160: standardised to 0.28% hypericin ● 3 350 mg (1050 mg) STEI 300: 0.5% hyperforin ● 3 300 mg (900 mg) WS5570: 0.3 Clinical evidence statements for St John’s wort compared with placebo138 Effect of treatment on efficacy outcomes There is some evidence suggesting that there is a clinically important difference favouring St John’s wort over placebo on increasing the likelihood of achieving a 50% reduction in symptoms of depression as measured by the HRSD in: ● the dataset as a whole (K 6139. 10. Schrader00. N 995. 5% hyperforin ● 3 300 mg (900 mg) WS5573: 0. 0. and data from up to 1629 participants were available from comparison with antidepressants.3% hypericin.495 mg total hypericin per tablet ● 900 mg LI 160 ● 4 200 mg (800 mg) LoHyp-57: drug extract ratio 5–7:1 ● 3 300 mg (900 mg) WS5572: drug extract ratio 2.Pharmacological interventions TCA-related antidepressants (Harrer94). Davidson02. all participants were aged 60 years and over.6 mg hypericin. Harrer99. Included studies described the following range of preparations: ● 2 150 mg (300 mg) at 0.3% hypericum ● 3 300 mg (900 mg) LI 160 720 to 960 mcg hypericin ● 2 250 mg (500 mg) ZE117: 0. Hangsen1996. It is very difficult to assess the exact content of the preparation of St John’s wort used in included studies so no study was excluded on grounds of inadequate dose.88) 137Davidson02 138The and Philipp99 are 3-arm trials. 95% CI.5–5:1.71 to 0. VanGurp02)137. forest plots can be found in Appendix 19c. All participants had either moderate or severe depression. 2 to 3% hyperforin ● 2 200 mg (500 mg) ZE117: 0. Schrader98) were removed from the meta-analysis to remove heterogeneity from the dataset.47 weeks).

389 . 95% CI.81.41). N 804. N 299. 95% CI. N ● severe depression (K 5 891. RR 0.22). –0.25).47 to –0. RR 0.22) 142. 141Three studies (Davidson02.87 to 1. in moderate depression there is some evidence suggesting that there is a clinically important difference favouring St John’s wort over placebo on reducing symptoms of depression by the end of treatment as measured by the HRSD (K 2. 95% CI. 142Ibid.79) severe depression (K 5140.96. 95% CI.72 to 1. There is insufficient evidence to determine if there is a clinically important difference between St John’s wort and placebo on increasing the likelihood of achieving remission by the end of treatment as measured by the HRSD (K 3. –1. Schrader98) were taken out of the meta-analysis to remove heterogeneity from the dataset. There is evidence suggesting that there is no clinically important difference between St John’s wort and placebo on reducing the likelihood of patients reporting adverse effects (K 7.74 to 1. RR 0. However. N 1127. –0. N 1612.22) 140Two studies (Davidson02. Acceptability and tolerability of treatment There is evidence suggesting that there is no clinically important difference between St John’s wort and placebo on reducing the likelihood of patients leaving treatment early for any reason (K 8. 95% CI.03. 95% CI.35.53 to 1.71. 0.89. 0.32 to 2. RR 0. N 898. There is insufficient evidence to determine if there is a clinically important difference between St John’s wort and placebo on reducing the likelihood of patients leaving treatment early due to adverse effects (K 5. Random effects SMD –0. 0.34. SMD –0. 0.64. 95% CI. SMD –0.51 to 0.13). N 1106. Random effects RR 0. but the size of this difference is unlikely to be of clinical importance in: ● the dataset as a whole (K 6141. 95% CI. N 1472.80. 0.2).47 to –0. Hangsen1996) were removed from the meta-analysis to remove heterogeneity from the dataset. There is evidence suggesting that there is a statistically significant difference favouring St John’s wort over placebo on reducing symptoms of depression by the end of treatment as measured by the HRSD. 143The forest plots can be found in Appendix 19c. 0. Random effects RR 1.4 Clinical evidence statements for St John’s wort compared with antidepressants143 Effect of treatment on efficacy outcomes There is evidence suggesting that there is no clinically important difference between St John’s wort and antidepressants on: ● increasing the likelihood of achieving a 50% reduction in symptoms of depression as measured by the HRSD (K 10.28 to –0.9.88. N 162.9). RR 0. 95% CI.1). 10. Hangsen1996. N 1031.72 to 0. 0. 95% CI.Pharmacological interventions ● ● moderate depression (K 1.

other than in moderate depression where it is better at achieving response and in severe depression where it is less effective than low-dose antidepressants in achieving response.17) ● reducing symptoms of depression by the end of treatment as measured by the HRSD (K 9.1).69.57 to 0.77. –0. However. Sub-analyses by antidepressant class and by antidepressant dose (therapeutic versus low dose) found similar results. N 1168. 95% CI. RR 1.6) ● reducing the likelihood of patients reporting adverse effects (K 8.95).9.26 to 0. 95% CI.44). RR 0. N 224. with fewer people leaving treatment early due to side effects and reporting adverse events. However. there is some evidence suggesting that there is a clinically important difference favouring St John’s wort over antidepressants given at therapeutic doses (K 5.47 to 1). 95% CI. Acceptability and tolerability of treatment With regard to reducing the likelihood of patients leaving treatment early for any reason.87 to 1.Pharmacological interventions ● increasing the likelihood of achieving remission by the end of treatment as measured by the HRSD (K 1. N 1011. 0. 0.75). St John’s wort appears as acceptable as placebo and more acceptable than antidepressants. and on reducing symptoms of depression in moderate depression. N 1629. N 521. 95% CI. A sub-analysis by severity found no difference in these results except for response rates in those with moderate depression: In moderate depression there is some evidence suggesting that there is a clinically important difference favouring St John’s wort over antidepressants on increasing the likelihood of achieving a 50% reduction in symptoms of depression as measured by the HRSD (K 3. There is strong evidence suggesting that there is a clinically important difference favouring St John’s wort over antidepressants on: ● reducing the likelihood of patients leaving treatment early due to side effects (K 10.13 to 0. RR 0. 1 to 1. 0.62 to 0.5 Clinical summary St John’s wort is more effective than placebo on achieving response in both moderate and severe depression. A sub-analysis combining severity and antidepressant dose also found similar results apart from for response rates in severe depression: In severe depression there is some evidence suggesting that there is a clinically important difference favouring low-dose antidepressants over St John’s wort on increasing the likelihood of achieving a 50% reduction in symptoms of depression as measured by the HRSD (K 4.65.39. RR 0. there is insufficient evidence to determine a difference between St John’s wort and either all antidepressants or low-dose antidepressants. 95% CI. N 481. 0. N 1358.2. 10. particularly TCAs. RR 1. RR 0. SMD –0. 95% CI.01.02. 390 . 95% CI. There appears to be no difference between St John’s wort and other antidepressants. 0.

6 10. Any wording changes have been made for clarification only. Costs were calculated from the societal 144The evidence for this recommendation has not been updated since the previous guideline. variation in the nature of preparations and potential serious interactions with other drugs (including oral contraceptives.1 Recommendation Although there is evidence that St John’s wort may be of benefit in mild or moderate depression. 2005b. Wade et al. Pharmacological companies producing the drugs under review were identified and contacted to provide/recommend unpublished or soon-to-be published studies in order to ensure up-to-date evidence was included in the evidence base for the guideline. Wade et al.2 Escitalopram and duloxetine Five industry-funded studies that assessed the cost effectiveness of escitalopram and duloxetine against various antidepressant comparators in the UK were included in the systematic review of economic literature (Benedicte et al. 2008). 2010. Wade and colleagues (2005a) investigated the cost effectiveness of escitalopram at a dose of 20 mg per day compared with citalopram at 40 mg per day in those with severe depression (MADRS 30) in primary and secondary care in the UK.10 10.. It incorporated effectiveness data derived from a study review and expert opinion. Data for response rates and other relevant inputs such as remission and discontinuation rates were derived from a 506-sample meta-analysis reporting at week 8. 391 .9. 2005a. This cost-effective analysis was reported to be an adaptation of models described in other studies such as Borghi and Guest (2000). Fernandez et al. 10. practitioners should: ● not prescribe or advise its use by people with depression because of uncertainty about appropriate doses. 2005.1 HEALTH ECONOMIC EVIDENCE Systematic literature review and economic considerations The systematic search of the economic literature undertaken for this guideline update identified nine studies.. Wade et al. these were then extrapolated to 6 months. 10.. anticoagulants and anticonvulsants) ● advise people with depression of the different potencies of the preparations available and of the potential serious interactions of St John’s wort with other drugs144.6. persistence of effect.Pharmacological interventions 10... Two unpublished evaluations submitted by pharmaceutical companies were also included.10. A decision tree with a 6-month time horizon was developed.9.10.

Escitalopram also fared better in terms of the effectiveness outcomes. expert opinion was also sought and unit costs were taken from published cost data for the UK. 1023) for citalopram. however. Escitalopram (10 to 20 mg daily) was compared with venlafaxine-XR (75 to 150 mg daily) and then generic citalopram (20 to 40 mg daily) over a 6-month time horizon from the perspective of the NHS and society. The price year was 2003. A decision tree representation was developed. dosage titration and switching of antidepressants as required. Conventional resource use directly related to treatment as well as treatment-emergent adverse events and attempted suicide were also included. The General Practice Research Database (GPRD) was searched for treatment pattern data. Escitalopram was demonstrated to be more effective and less costly. Univariate sensitivity analysis and Monte Carlo simulations were conducted to evaluate the effect of uncertainty. In the comparison between escitalopram 392 .Pharmacological interventions perspective as well as from that of the NHS and reported in 2003 pound sterling. GP visits and psychiatrist visits. The direct health service costs included in the economic evaluation were drugs. An Austrian costeffectiveness model (Hemels et al.7% (50.8 to 51. For example. The model encompassed remission. Lost productivity costs due to absenteeism from work were calculated using the human capital approach. which examined the cost effectiveness of three drug therapies for the treatment of depression in primary care. Multivariate sensitivity analysis further demonstrated that escitalopram was dominant at all ranges of probabilities tested in more than 99% of simulations.7% (45. based on mean market wages for 2003.5%) compared with 48. defined as a patient in remission (MADRS 12 at week 24). There are many commonly used drugs for depression and other comparators from other drug classes may have been relevant for analysis and their inclusion would possibly have been more informative. the expected total cost per patient was £422 (£404 to £441) for escitalopram and £454 (£436 to £471) for citalopram. This study is deemed to be of good quality. while the secondary outcome measure was first-line success (that is. Because of an absence of relevant head-to-head studies. remission without switching drug treatment). The primary outcome measure was patient treated successfully. 807) for escitalopram and £933 (95% CI 850. The summary benefit measure was the overall success rate and this was estimated using the decision model. referral to secondary care. Another study by Wade and colleagues (2005b) was reviewed. Cost estimates for the majority of the resources used were derived from national published studies. When escitalopram was compared with citalopram from the NHS perspective the cost per successfully treated patient was £732 (95% CI 665. depression is a chronic illness and a 6-month time horizon may well be too short to capture all costs and benefits. The clinical evidence came from a metaanalysis of four studies (n 1472) and head-to-head clinical trials.. overall success was 53.3 to 57. treatment failure. Wade and colleagues (2005a) concluded that escitalopram was a cost saving alternative to citalopram for the treatment of people with severe depression in the UK despite the price of escitalopram being higher than other generic drugs. two separate analyses were run. and therefore escitalopram dominated citalopram.7%) for citalopram. From the NHS perspective. Cost savings were shown from both perspectives. 2004) was adapted for the UK by Wade and colleagues (2005b).

An 8-week follow-up is quite short for a depression-related study and. drug dosages and resource utilisation differed in both.8 to 12. The analysis of efficacy data was based on the basis of treatment completers only. However. The direct costs for the average patient were reported to be 40% higher for venlafaxine-XR than for escitalopram. the comparison with venlafaxine was sensitive to changes in parameters such as remission rates and relapse rates used in the model. However. Spain and the UK. in the first duloxetine was compared with SSRIs. Quality of life (QoL) is an important dimension in the depression spectrum and the impact of the interventions under review on QoL may have been informative. the cost per successfully treated patient was £546 (95% CI 481. 677) for citalopram. study was unpublished during the development of this guideline update. Costing was undertaken prospectively on the same patient sample. Germany.6 to 12. multinational145 RCT with 8-week follow-up (n 293). 618) for escitalopram and £607 (95% CI 542.4 for escitalopram-treated patients (p 0. An indirect comparison analysis could have been conducted had there been relevant head-to-head trials published. The primary health outcome was quality of life measured on the Quality of Life Depression Scale (QLDS). which were more expensive and less effective. It described an economic evaluation of duloxetine in comparison with SSRIs.Pharmacological interventions and venlafaxine. CEACs were not produced because there were no significant differences in efficacy. Sensitivity analysis showed robust findings for the analysis between escitalopram and citalopram.01) and from 18. ICERs were not calculated because escitalopram was found to always dominate both citalopram and venlafaxine. 393 . The study by Benedicte and colleagues (2010)146 was also reviewed. Fernandez and colleagues (2005) aimed to assess the cost effectiveness of escitalopram (10 to 20 mg/day) compared with venlafaxine-XR (75 to 150 mg/day) in UK primary care patients with depression. The second analysis set in secondary care compared duloxetine with venlafaxine and mirtazapine in patients with severe depression (HAMD-17 25). 146This Finland. The authors concluded that escitalopram is as effective as venlafaxine-XR on the treatment of depression and may be associated with lower costs from both perspectives. Limited details of the effectiveness study were reported making it difficult to assess the study quality or validity.1 for venlafaxine-treated patients (p 0. Mean QLDS scores decreased in both groups: from 18. the authors argue that an indirect comparison would not have changed the conclusions of the analysis. Efficacy data. venlafaxine and mirtazapine in patients with moderate to severe depression (HAMD-17 19) in primary care. as a result. France. The effectiveness data were derived from a double-blind. Fernandez and colleagues (2005) acknowledged that larger sample sizes are required to increase the power of performed tests and to enable the detection of differences in costs between escitalopram and venlafaxine-XR. venlafaxine-XR and mirtazapine in primary and secondary care settings in Scotland. No statistically significant differences were observed between the groups.01). long-term costs and benefits may not have been captured. The perspectives of the NHS and society were adopted. Two analyses were conducted. The results showed escitalopram to be less costly and equally effective as venlafaxine-XR. The perspective adopted 145Denmark.

10.400/QALY gained. The perspective adopted in this study is not that of the health services and is therefore less useful for those making decisions on behalf of health services. duloxetine relapse. Because of the marked differences in healthcare systems there would be differences in healthcare resource use costs and the relative economic burden of sick leave. literature and UK practicing GPs. The study limitations considered that efficacy data for SSRIs had been collected from other duloxetine trials and for mirtazapine from a single old meta-analysis of limited quality. In the primary care setting. Indirect costs because of sick leave accounted for two-thirds of the total costs. Resource use estimates over this time were sourced from the health economics assessment questionnaires taken alongside the trial. The secondary care scenario was less sensitive to changes. Direct medical costs consisted of all outpatient and inpatient visits and drug costs. The clinical effectiveness parameters were from published and unpublished RCT data.3 Selective serotonin reuptake inhibitors.Pharmacological interventions was that of the NHS. The study time horizon was 24 weeks. Wade and colleagues (2008) evaluated the cost effectiveness of escitalopram and duloxetine in the treatment of patients with depression in an outpatient setting. The results showed that over the study period escitalopram was associated with significant cost savings compared with duloxetine (£1127 versus £2. TCAs and lofepramine (a newer TCA which is safer in overdose) in the treatment of depression in adult patients 394 . This analysis was carried out alongside a double-blind. multisite randomised study. Escitalopram dominated duloxetine in the primary analysis (that is. The primary effectiveness outcome of the analysis was the Sheehan Disability Scale (SDS) score. The short time horizon modelled may not capture all the costs and benefits of the drugs for the treatment of depression. Resource use estimates were sourced from the Scottish Psychiatrists’ Panel. when assessed with the SDS scale). 10. remission and response rates). Escitalopram also resulted in significantly lower sick leave duration compared with duloxetine (31 versus 62 days). Resource use data were collected from a small physician panel that is not considered to be a good source of such evidence. when compared with SSRIs and mirtazapine. which limits the generalisability of the results to the UK. The societal perspective was adopted and results were reported in 2006 UK pound sterling. Duloxetine also dominated venlafaxine and mirtazapine in the secondary care setting. tricyclic antidepressants and lofepramine One study that assessed the cost effectiveness of SSRIs. The cost effectiveness results in the primary care setting were sensitive to changes in efficacy parameters (that is.001 total/patient cost respectively). duloxetine produced additional benefits at higher costs leading to ICERs of £6. The authors acknowledged the risk of bias given the problems of comparability of trial populations. It dominated venlafaxine in this setting. The main outcome of the model was QALYs. other clinical study data and expert interviews. This study was conducted in several countries in addition to the UK.300/QALY and £2.

randomised. The CEACs suggested that. openlabel. The study adopted the perspective of the health service.000. although the probability of this did not rise above 0. It also drew from a population from several centres across the UK. the number of psychiatric consultations with GPs and the percentage of patients who completed 6 weeks’ 395 . The costing was carried out prospectively on the same sample (n 327) of patients. a panel of ten GPs and three psychiatrists.Pharmacological interventions in the UK was included in the systematic review of economic literature (Kendrick et al. The cost-effectiveness planes for each comparison included points in all four quadrants reflecting statistically non-significant differences in outcomes and costs. community psychiatric nurse and community mental health team visits. This analysis was based on a trial that was well described and reflected usual practice. The differences in the total costs did not reach statistical significance either. antidepressant and concomitant medication. Cost-effectiveness planes and CEACs were computed to illustrate the uncertainty around the estimates. The primary clinical measure was the number of weeks free from depression (HADS-D 8). Mirtazapine was observed to be cheaper and more effective than amitriptyline and therefore dominated amitriptyline. as well as the costs related to antidepressant discontinuation.4 Mirtazapine and venlafaxine Two industry funded UK based studies compared mirtazapine to older agents such as TCAs and SSRIs (Borghi & Guest. Borghi and Guest (2000) aimed to determine the cost effectiveness of mirtazapine compared with amitriptyline and fluoxetine in the treatment of moderate and severe depression in the UK. 2000. clinical trial in primary care from the perspective of the health service. Loss to follow-up approaching 50% over 12 months further limited the power. The length of follow-up was 12 months. Mirtazapine’s cost effectiveness relative to amitriptyline was sensitive to the cost of managing adverse events. The measure of benefit used was the proportion of successfully treated patients. No statistically significant differences between the groups were observed in this measure. for values placed on an additional QALY of over £5.. and attendance at day wards. Direct costs included costs of hospitalisation. 10. visits to GPs and psychiatrists. Six months’ treatment with mirtazapine compared with fluoxetine increased the proportion of successfully treated patients by 22% at a net additional cost to the NHS of £27 per patient. SSRIs were likely to be most cost effective. A limitation of the study was the failure to recruit the desired number of patients thereby reducing the study’s power to detect differences in effectiveness and costs. and literature. 2006b). The estimation of quantities and costs was based on actual data.. The price year was 1997/1998.6. Mirtazapine’s cost effectiveness relative to fluoxetine was sensitive to the cost of managing patients who discontinue antidepressant treatment. Effectiveness data were derived from a literature review and also from a panel of GPs and psychiatrists. determined by the HAMD-17 score (7 or less). Romeo et al. This trial provided effectiveness and costing data. 2004). The study was carried out alongside a prospective.10. which was representative of the wider UK population.

Another limitation was that resource use data were obtained from interviews with a panel of experts. this is not considered to be ideal. they were not combined in the form of ICERs. which was conducted prospectively. This model was adapted with country specific estimates from a clinical management analysis. If the willingness-to-pay increased to £1000. were obtained from a subgroup of patients participating in the trial (treatment completers only). Romeo and colleagues (2004) concluded that compared with paroxetine. The potential limitations are that the analysis may be subject to potential selection bias. The model data were obtained from an RCT. Compared with paroxetine. The authors recommend an update of the model when longer-term data are available. it was reported that patients excluded from the subgroup did not differ from the patients included in terms of baseline characteristics. There were no significant differences in costs and effects on the primary outcome measure. A decision analytic model with a 6-month time horizon was developed. this probability rose to 89%. Sensitivity analysis revealed that if society were willing to pay nothing for a point improvement in depressive syndromes. SSRIs and TCAs in acute depression. The study was conducted in general practices in Scotland and had a 24-week follow-up. In addition. improvement in quality of life was shown to be significantly higher with mirtazapine than with paroxetine. Both antidepressants were efficacious for 24 weeks of treatment in depressed primary care patients. when considering improvements in quality of life following the administration of these two agents. The subgroup used consisted of treatment completers only. No further statistical analyses. Nevertheless. Romeo and colleagues (2004) compared the cost effectiveness of 30 to 45 mg/day mirtazapine with 20 to 30 mg/day paroxetine for patients with depression treated in primary care. therefore. The primary measure of cost effectiveness was the incremental cost per responder. A significant limitation of this study was that 6-week data comparing mirtazapine with fluoxetine was extrapolated to 6 months using assumptions derived from published literature due to the lack of available comparison data at the time of the study. were undertaken. Costs were reported from the NHS and societal perspectives.Pharmacological interventions treatment with mirtazapine and achieved a 50% reduction in the HAMD-17 score. However. Doyle and colleagues (2001) described a multinational pharmacoeconomic evaluation which compared the cost effectiveness of venlafaxine. to account for potential biases and confounding factors. patients with a 50% decrease in the HAMD-17 score) and the change in QLDS score (from baseline) at the 24-week end point to capture change in quality of life. Effectiveness outcomes were reported in the form of number of HAMD responders (that is. The effectiveness data and costing. However. meta-analytic rates and two published meta-analyses and a resource 396 . there was an 80% probability that mirtazapine would be more cost effective than paroxetine. there were no significant differences in the benefits between the two groups when the number of HAMD responders was the outcome considered. mirtazapine might be a cost-effective treatment choice for depression in a primary care setting. mirtazapine was associated with greater improvements in quality of life. it can be inferred that mirtazapine should be considered the treatment of choice. These results were robust under all scenarios examined in the sensitivity analysis.

they show no robust clinically important superiority in terms of effectiveness. (In the previous guideline. Because of the marked differences in healthcare systems there would be differences in healthcare resource use patterns and patient variations. the lack of any greater effect than older drugs makes the added cost potentially not worthwhile (see Table 83). a better tolerated drug may also result in cost savings because of the potential decrease in adverse event related healthcare resource use.Pharmacological interventions valuation of treatment costs by local health economists in each country. Such an evaluation could inform future guideline recommendations. The findings from the health economic evidence highlighted the need for de novo economic modelling for this guideline (see Section 10. Most SSRIs are off patent and available in generic form. This study was conducted in several countries in addition to the UK. pharmacoeconomic evidence suggested that SSRIs were more cost effective than TCAs for the first-line treatment of depression. There is also a weak trend that reflects that SSRIs may be more cost effective than TCAs. Cost effectiveness was determined using the expected values for both a successful outcome and a composite measure of outcome termed ‘symptom-free days’. side effect profile. The guideline update recommends that normally an SSRI should be prescribed because they are as effective as other antidepressants. 10.) In the previous guideline. that is. pharmacoeconomic evidence suggested that venlafaxine was more cost effective than SSRIs. The short time horizon modelled may not capture all the costs and benefits of the drugs for the treatment of depression. 397 . are better tolerated and are less likely to be discontinued because of side effects. however. In the case of newer drugs.5 Summary of health economic evidence The pharmacoeconomic evidence (much of it industry funded) presented above suggests that escitalopram is better in terms of costs and benefits compared with some of the antidepressants. which limits the generalisability of the results to the UK. These findings were explored with sensitivity analysis. cost of drug and patient choice. It is evident that the nature of the current pharmacoeconomic data is piecemeal – no study compares all the relevant antidepressants drugs in a single evaluation. Venlafaxine dominated the other two options since its expected total health service costs were the lowest and it was more effective in terms of both success rate and symptom free days. many factors should be taken into consideration for example. the clinical evidence review at the time highlighted that the clinical estimates used in the economic studies of the drugs compared were inconsistent with the results of the NCCMH clinical evidence review.12). when making a treatment decision regarding the use of an antidepressant. Therefore. Therefore an opportunity cost approach was adopted and primary care costs of the different antidepressants were considered alongside the clinical evidence.10. Additionally. The updated meta-analyses of clinical evidence in this guideline points to similar levels of effectiveness across the antidepressants reviewed. clinical history.

73 Venlafaxine-XR 100 mg £7.24 Weekly cost £3. 56-tab pack £31. 2009) 60 mg 28-cap pack £27. 28-tab pack £1. They reported that ‘mirtazapine.Pharmacological interventions Table 83: Drug acquisition costs Drug Escitalopram Average daily quantities unit 10 mg Unit cost (BNF 56. 2009). 2008) 75 mg 28-cap pack £23. escitalopram.31 *Based on the Electronic Drug Tariff as of 23 May 2009 (NHS. 10. maintenance therapy or switching and sequencing patterns were identified in the UK setting. fluvoxamine.91 Efexor® XL (Wyeth Pharmaceuticals.80 £5. venlafaxine. fluoxetine. 2009) 10 mg (scored). milnaciran. citalopram.31 £0. This was a network meta-analysis which looked at the comparative evidence from RCTs for 12 antidepressants using both direct and indirect methods.26 £6. No new pharmacoeconomic evidence on relapse prevention. escitalopram. 28-tab pack £14. 398 . this provides a valid way of comparing individual drugs taking into account results against other drugs in the ‘network’ as well as being able to compare drugs in the absence of head-to-head RCT evidence. 1·30 and 1·27. The authors demonstrated that sertraline. duloxetine.93 – £0. September 2008) Cipralex® (Lundbeck.11 NETWORK META-ANALYSIS OF NEWER ANTIDEPRESSANTS A review by Cipriani and colleagues (2009) was published at the end of the guideline development process and was considered by the GDG in view of its method and potential importance. net price 28-tab pack £1. mirtazapine and venlafaxine performed well in terms of efficacy and tolerability compared with the other antidepressants reviewed (bupropion. Business Services Authority.72 Not available 20 mg.61* Cymbalta® (Eli Lilly. and sertraline were significantly more efficacious than duloxetine (odds ratio [OR] 1·39.33 Duloxetine Agomelatine Citalopram Sertraline 60 mg Not available 20 mg 50 mg 50 mg. paroxetine and reboxetine). 1·33.41 Non-proprietary 75 mg.

with the comparators moving up the ranking slightly. respectively). discontinuation symptoms and social functioning (Cipriani et al. this uncertainty aside.12. paroxetine ([OR] 1·35. fluoxetine ([OR] 1·37. Sertraline might be the best choice when starting treatment for moderate to severe major depression in adults because it has the most favourable balance between benefits. 1·30. the size of the efficacy effect when translated from the odds ratio reported in the study to an absolute risk is small.10. toxic effects. Cipriani and colleagues’ (2009) analysis found that the cumulative probability of being among the four best treatments became slightly smaller for those drugs in trials that were sponsored by the marketing company. 1·27. the analysis was limited to response rates (some of which were imputed) and this outcome measure may provide a less conservative measure of effect than the other commonly used measures (remission and continuous data). The analysis was based on efficacy data (response rates) and dropout rates using data from 117 trials (about 26. paroxetine.. Fourth. and reboxetine ([OR] 2·03. Fifth. 1·95. 1·30. leading to significantly fewer discontinuations than did duloxetine. This effect. and 1·85. 10. such as evidence of side effects. acceptability. and 1·25. the size of the tolerability effect is small when translated from an odds ratio to an absolute risk. highlights the difficulty in excluding potential confounds. For example. while likely to be small. They did not consider other potentially important factors. and 1·22. and 1·27. Finally.Pharmacological interventions respectively). Second.1 As described in Section 10. There are some methodological aspects of the study that are important to consider. Reboxetine was significantly less efficacious than all of the other antidepressants tested. as the method of analysis may favour the drug with fewer dropouts. Third. for example with drugs like reboxetine and escitalopram. and venlafaxine’.7% for sertraline versus fluoxetine. The credibility interval encompassed much higher values. reboxetine. fluvoxamine. Escitalopram and sertraline showed the best profile of acceptability. fluvoxamine ([OR] 1·41.12 ECONOMIC MODEL FOR THE COST EFFECTIVENESS OF PHARMACOLOGICAL INTERVENTIONS FOR PEOPLE WITH DEPRESSION Introduction 10. and acquisition cost’. respectively). it is not clear to what degree differential dropout rates may have influenced the relative efficacy. the systematic search of economic literature identified a number of studies on pharmacological treatments for the management of depression 399 . 2009). Cipriani and colleagues (2009) concluded that ‘clinically important differences exist between commonly prescribed antidepressants for both efficacy and acceptability in favour of escitalopram and sertraline. total dropouts may not be an accurate way to assess tolerability and usually only half of dropouts are attributed to adverse effects. respectively). respectively). First. 1·28. The credibility interval again encompassed much higher values. it is about 2. 1·32. This adds uncertainty to the analysis.000 participants). 1·35. 1·89. assuming a dropout rate of 28% on fluoxetine from Table 38 in Cipriani and colleagues (2009) (number needed to harm [NNH] 37).

Two separate analyses were conducted for hypothetical cohorts of 100 patients with either moderate or severe depression. sertraline and venlafaxine. People who discontinued from initial antidepressant treatment were assumed to receive various levels of care for their depression. fluoxetine.12. Switching to second-line antidepressants was not considered for those patients who discontinued their first-line antidepressant treatment or who did not respond to treatment. reboxetine. The studies were characterised by varying quality in the methods employed. Patients who responded to initial treatment received 6 months of maintenance therapy and then were assumed to either relapse or enter remission. The objective of economic modelling was to explore the relative cost effectiveness of antidepressants for people with depression in the current UK clinical setting. including no care. The analysis was based on 117 studies including 25. These findings highlighted the need for de novo economic modelling for this guideline. from the network meta-analysis is acknowledged as a potential limitation for the economic analysis.11.. The model constructed for the economic analysis of combination therapy versus antidepressant treatment in Section 8. The exclusion of other categories of antidepressants. each assessing the relative cost effectiveness of the ten 400 . milnacipran. Some of these people were assumed to clinically improve.928 participants randomly assigned to 12 different new-generation antidepressants. The time horizon of the analysis was 14 months. such as TCAs and MAOIs.2 Methods Interventions assessed The choice of interventions assessed in the model was determined by the antidepressants included in the network meta-analysis by Cipriani and colleagues (2009). patients were initiated on a specific antidepressant and either continued or discontinued treatment. Patients continuing their initial antidepressant treatment either responded or did not respond.9 was adapted for this analysis. The number of antidepressants assessed in this literature was limited and did not include the whole range of drugs available in the UK for the treatment of people with depression. mirtazapine. Model structure A pragmatic decision analytical model was constructed using Microsoft Excel XP. plus 12-month follow-up. as described in Section 10. and then either relapse or enter remission. 10. duloxetine. escitalopram. These included bupropion. fluvoxamine. incorporating the results of a recently published network metaanalysis (Cipriani et al. For the economic model. this consisted of 2 months of treatment. paroxetine. The remaining ten antidepressants were assessed in the economic model. Within the antidepressant model. citalopram. bupropion and milnacipran were excluded from the analysis because bupropion is not currently licensed as a treatment for depression and milnacipran does not currently have a licence for treatment in the UK. reflecting the time point at which the clinical efficacy and acceptability parameters reported in Cipriani and colleagues (2009) were measured. for which relapse data were available. 2009).Pharmacological interventions in the UK.

Acute treatment was defined as 8 weeks of treatment for both efficacy (response) and discontinuation (drop out) analyses. only RCTs that compared the following 12 new-generation antidepressants were considered: bupropion. Efficacy and discontinuation data Overview of methods used by Cipriani and colleagues (2009) In summary. milnacipran. A schematic diagram of the economic model is presented in Figure 9. paroxetine. fluvoxamine.Pharmacological interventions antidepressants assessed. escitalopram. The measure of outcome was the QALY. If 8-week data were not available. duloxetine. Response was defined as the proportion of patients who had a reduction of at least 50% from the baseline score on the HRSD or MADRS or who scored much improved or very much improved on the CGI scale at 8 weeks. Figure 9: Schematic diagram of the economic model structure Relapse Response Continue antidepressant Remission No Response Patients with moderate/severe depression Relapse Response Remission Discontinue antidepressant No Response 401 . sertraline and venlafaxine as monotherapy in the acute-phase treatment of adults with depression. fluoxetine. Costs and outcomes considered in the analysis The analysis adopted the NHS and PSS perspective. Treatment discontinuation was defined as the number of patients who stopped treatment early for any reason during the first 8 weeks. mirtazapine. reboxetine. data ranging between 6 and 12 weeks were used. citalopram.

the probabilities for discontinuation and response were applied over a period of 2 months in the economic model. The rate of relapse for the pharmacotherapy arm over 12 months was 0. it was assumed that they would either relapse or enter remission. Fluoxetine was used as the reference drug. and OR(FL.55 and was applied irrespective of initial antidepressant treatment. a random-effects model within a Bayesian framework using Markov chain Monte Carlo methods was used.AD) OddsFL/OddsAD (3) PAD OddsAD/(1 OddsAD) OddsFL and PFL are the odds and probability (of relapse or dropping out) for fluoxetine at 8 weeks.AD) is the odds ratio of fluoxetine versus each antidepressant (of relapse or dropping out) at 8 weeks. For the network meta-analysis. are presented in Table 85. OddsAD and PAD are the odds and probability (of relapse or dropping out) for each of the other antidepressants considered at 8 weeks. Because both response and dropout rates referred to an 8-week period. The odds ratios reported for fluoxetine versus each of the other antidepressants were converted into probabilities (response and dropouts) for each antidepressant considered in the economic model using the following formulae: (1) OddsFL PFL/(1 PFL) OddsAD OddsFL/OR(FL. and it had been consistently used as the reference drug among the different pair-wise comparisons in the RCTs considered in the network meta-analysis. Two of the trials had three arms and compared fluoxetine with paroxetine and sertraline. The comparative efficacy and acceptability among the 12 antidepressants was shown in terms of odds ratios of each antidepressant versus fluoxetine. 402 . Results were reported as odds ratios for all pairs of antidepressants that were considered in the network meta-analysis. The rate of relapse for these patients was taken from the guideline meta-analysis of relapse over 12-month follow-up for the economic model of combination therapy compared with antidepressant treatment (see Section 8. and therefore have been double-counted at the estimation of probabilities on response and dropping out for fluoxetine because it was not possible to identify and isolate respective data coming from these two RCTs. along with their 95% credible intervals.Pharmacological interventions Responders to treatment in each trial were calculated on an intention-to-treat basis.AD) (2) OR(FL. Outcomes were imputed for missing participants assuming that they did not respond to treatment. The probabilities for response and discontinuation for each antidepressant over 8 weeks. because it was the first among the 12 antidepressants to be marketed in Europe and the US. The data on fluoxetine from these two trials were reported twice.9). Estimation of response and discontinuation rates in the economic model The efficacy and acceptability results from the network meta-analysis by Cipriani and colleagues (2009) are summarised in Table 84. The probabilities for fluoxetine were estimated based on data reported for 54 RCTs considered in the network meta-analysis that included fluoxetine in one of their arms. Other model clinical input parameters For patients who responded to initial antidepressant treatment after 2 months.

81–1.91–1.53–0. 1984).97 (0. again it was assumed that a proportion would relapse and the remaining patients would enter remission.14 (0.31) 0.80 (0. expressed as odds ratios (OR) of fluoxetine versus each of the antidepressants assessed (taken from Cipriani et al. these rates were applied to all patient cohorts irrespective of initial antidepressant treatment. All remaining patients in the model of those who responded to initial antidepressant treatment (that is. OR lower than 1 favours fluoxetine.86–1. Of those patients who improved following discontinuation.73 (0.74–1.64–1.93)* 0.30) 0.11) 0.76 (0.68–0.77–1.88)* 0. OR higher than 1 favours fluoxetine.81–1.69–0.93 (0.82 (0. For patients who discontinued their initial antidepressant treatment at 8 weeks.92)* 1.10) 1.21) 0.62–1.84 (0. QoL weights estimated in a study of patients with depression were used (Sapin et al.12 (0.70 (0..89)* 1. For acceptability.78 (0.. those who did not relapse) were assumed to enter remission.05.90)* Acceptability (dropout rate) OR (95% credible interval) 1.65–0.36) 1.91 (0.36) 0.9 for further details).32) 0.98 (0.48 (1. *p 0.81–1.37) 0. Utility weights used in the economic analysis are presented in Table 86. either spontaneously or following treatment (according to ‘response’ as defined in Cipriani et al. a proportion (20%) would improve from their baseline health state.09) Credible interval.99–1.12) 1.94 (0.92–1..11 (0.27) 0.08) 1.76–1. it was assumed that rather than remaining moderately or severely depressed.67 based on a study of patients who were not receiving maintenance therapy at 12 months (Murphy et al.05) 0.79–1. For efficacy.19 (0.96–1.Pharmacological interventions Table 84: Efficacy (expressed as response rate) and acceptability (reflected in dropout rates) of antidepressants. 2009) Efficacy (response rate) OR (95% credible interval) Bupropion Citalopram Duloxetine Escitalopram Fluvoxamine Milnacipran Mirtazapine Paroxetine Reboxetine Sertraline Venlafaxine 0.97 (0.90)* 0.69–1.77–1.16–1.60–0.07) 0. The rate of relapse for these patients was assumed to be 0.02 (0.. 2004) (see Section 8. 2009).99 (0.91 (0. Estimation of quality-adjusted life-years To calculate QALYs. 403 . Again.44) 0.01 (0.

45 (0.24–0.64) Acceptability (dropout rate) Probability (95% credible interval) 0.28 (0.26–0.29) 0.62 (0.33 (0.65) 0.22–0.87) 0.32 (0.55 (0.08–0.50–0.37) 0.83–0.30) 0. Resource use and unit costs An NHS and PSS perspective was taken for the analysis based on current NICE guidance (NICE.33) 0.58–0.58–0.29–0.28) 0.48–0.65–0.72 (0.38) 0.39–0.57 (0.54 (0.29–0.Pharmacological interventions Table 85: Probabilities estimated for use in the economic model (adapted from Cipriani et al.30 (0.29 (0.29) 0.33) 0.79) 0.51) 0.66) .58–0.28 (0.42) 0.57–0. only direct health and social care costs were considered in the model.60 (0.62) 0.64) 0..38) 0.60) 0.60) 0.26–0. Table 86: Quality-of-life weights utilised in the economic model Health state Moderate depression Severe depression Response with remission Response without remission No response (following treatment) 404 Quality-of-life weight (95% CI) 0.54–0.22–0.85 (0.26 (0.53–0. monitoring costs relating to consultations with psychiatrists and GP visits.31 (0.25 (0.56) compound) Citalopram Duloxetine Escitalopram Fluvoxamine Mirtazapine Paroxetine Reboxetine Sertraline Venlafaxine 0.52–0. Costs included drug acquisition costs.15 (0.49–0.21–0.59) 0.61 (0. 2009) Efficacy (response rate) Probability (95% credible interval) Fluoxetine (reference 0.55 (0.67) 0.26–0. as well as other health and social care costs associated with the care of people with depression who discontinued treatment.32) Note: Bupropion and milnacipran excluded from economic analysis.63 (0. 2008b).22) 0.58 (0.27–0. Therefore.27–0.55 (0.24 (0.35 (0.

The daily dosage of all ten antidepressant drugs was informed by the midpoint of the range of daily dosages presented in Cipriani and colleagues (2009) and by the BNF. For both costs and outcomes. the maintenance therapy period consists of one GP/specialist visit every 2 months.9). The total antidepressant treatment costs including patient monitoring are presented in Table 88. Business Services Authority. Other healthcare costs It was assumed that patients with moderate or severe depression would require additional subsequent mental health and social care resources if they discontinued their initial therapy. with the exception of the cost of venlafaxine which was obtained from the Electronic Drug Tariff (NHS. did not respond to their initial antidepressant treatment at 8 weeks. For all patients. Unit costs were obtained from a variety of sources including the BNF (British Medical Association and the Royal Pharmaceutical Society of Great Britain. 2008) and the PSSRU (Curtis. It was assumed that patients with moderate or severe depression would both receive the same average daily dosage. Costs were calculated by combining relevant resource use estimates with national unit costs. 2009). 2008). 2009) because this antidepressant has recently become available in generic form but BNF 56 has not captured this information. it was estimated that patient monitoring in both primary and secondary care consists of two fortnightly visits in the first month followed by one visit in the second month. The average daily dosages and the drug acquisition costs are presented in Table 87. Based on the same assumptions used in the combination therapy versus antidepressant treatment model (see Section 8. 2009). It was assumed that all patients who did not discontinue and responded to their initial treatment after 8 weeks would continue to receive maintenance antidepressant treatment at the same dose over a further 6 months in the model. or responded to treatment but relapsed at a later stage. According to the expert opinion of the GDG. the total costs of antidepressants were calculated over the 8 weeks of initial therapy. 2009). all patients with moderate depression and 50% of patients with severe depression would receive standard GP care while the remaining 50% of patients with severe depression would receive specialist mental health outpatient care. All costs were based on 2008 prices and were inflated where necessary using the Hospital and Community Health Services Pay and Prices Indices (Curtis. Drug acquisition costs Drug acquisition costs were taken from BNF 56 (British Medical Association and the Royal Pharmaceutical Society of Great Britain. The unit costs of a GP consultation (£36) and a mental health outpatient consultation (£130) were both taken from the latest PSSRU estimates (Curtis. or 405 . no discounting was applied given the short time horizon of the model (14 months).Pharmacological interventions or did not respond to treatment. Monitoring costs All patients receiving antidepressant treatment were assumed to be actively monitored either in primary or secondary care during both the initial treatment period and the maintenance treatment period.

32 £3. 28-tab £1.61* *Based on the Electronic Drug Tariff as of 23 May 2009 (NHS. it was not possible to explore uncertainty around key parameters used in the model. 30-tab Non-proprietary 100 mg. People who relapsed over the 12 months following initial therapy were assumed to relapse in the middle of this period. 30-tab Non-proprietary 30 mg. cost estimates and health state 406 . 28-tab Non-proprietary 20 mg.Pharmacological interventions Table 87: Average daily dosages and acquisition costs of antidepressant drugs included in the economic model Antidepressant Citalopram Duloxetine Escitalopram Fluoxetine Fluvoxamine Mirtazapine Paroxetine Reboxetine Sertraline Venlafaxine Average daily dosage 40 mg 60 mg 10 mg 40 mg 100 mg 30 mg 40 mg 8 mg 100 mg 150 mg Unit cost (BNF 56. responded to therapy but relapsed at a later stage. 56-tab £1. Sensitivity analysis Because of time constraints. monthly mental health and social care cost estimates (£180 per month) were estimated from a study that calculated annual mental health and social care costs based on responses from the UK psychiatric morbidity survey (McCrone et al. 2008). Based on the same assumptions used in the combination therapy versus antidepressant treatment model (see Section 8. 60-tab £18. 28-tab Non-proprietary 75 mg. 28-tab Cymbalta 60 mg.46 £27.46 £8.80 £31. 28-tab Cipralex 10 mg.14 £2. For patients who responded to initial treatment and did not relapse during follow-up..91 £1. 2008) Non-proprietary 40 mg.92 Non-proprietary 20 mg.91 Non-proprietary 100 mg. For both dropouts and non-responders. that is. it was assumed that no further additional treatment or mental health and social care resources beyond the 6-month maintenance period were required. 30-tab Edronax 4 mg.9). it was assumed that these costs were incurred over the 12 months following initial antidepressant treatment. at 6 months.72 £14. Therefore they were assumed to incur these mental health and social care costs for 6 months at the end of the maintenance therapy period. 2009). including resource use. These total subsequent mental health care costs are presented in Table 88. Business Services Authority.

Full access to this dataset is necessary in order to maintain the correlation between the posterior estimates when running the probabilistic analysis.Pharmacological interventions Table 88: Total healthcare costs applied in the economic model Antidepressant Total antidepressant treatment Total antidepressant treatment costs-MODERATE DEPRESSION costs-SEVERE DEPRESSION Initial treatment (8 weeks) Citalopram Duloxetine Escitalopram Fluoxetine Fluvoxamine Mirtazapine Paroxetine Reboxetine Sertraline Venlafaxine £111 £168 £140 £112 £125 £115 £120 £146 £112 £177 Maintenance treatment (6 months) £118 £289 £205 £120 £159 £128 £144 £223 £120 £314 Initial treatment (8 weeks) £252 £309 £281 £253 £266 £256 £261 £287 £253 £318 Maintenance treatment (6 months) £259 £430 £346 £261 £300 £269 £285 £364 £261 £455 Subsequent health states No response/dropout (12 months) Relapse (6 months) Mental health and social care costs £2. any options that are more expensive than options that are higher in ranking are dominated (because they are also less effective) and excluded from 407 . 10. Furthermore. Relative cost effectiveness between alternative treatment options is estimated using incremental analysis: all options are first ranked from the most to the least effective.12. probabilistic sensitivity analysis was not possible because this required full access to the posterior estimates recorded within every iteration of the network meta-analysis undertaken by Cipriani and colleagues (2009).3 Data analysis and presentation of the results A deterministic analysis was undertaken.160 £1080 utilities. Deterministic sensitivity analysis was only carried out on the upper and lower 95% credible intervals around the response and dropout probabilities (see Table 85). where data are analysed as point estimates. results are presented as mean total costs and QALYs associated with each treatment option assessed.

Subsequently. After having excluded cases of dominance and extended dominance. sertraline is the second most cost-effective option after mirtazapine and escitalopram is the third most cost-effective option. with results being statistically significant for the first four of them. escitalopram. 2008b). starting from the most to the least effective. Therefore. resulting in the lowest number of QALYs and the highest costs. Full results of the deterministic analysis for both moderate and severe depression are presented in Table 89. Estimation of such a ratio allows consideration of whether the additional benefit is worth the additional cost when choosing one treatment option over another. the rankings of antidepressants in terms of cost-effectiveness are very similar to the ranking of antidepressants in terms of efficacy. for patients with either moderate or severe depression. In their analysis. the next option would be escitalopram or sertraline because venlafaxine is dominated by escitalopram and the remaining antidepressants are dominated by sertraline. The treatment option with the highest ICER below the cost-effectiveness threshold is the most cost effective option. producing the highest number of QALYs and the lowest costs among all drugs assessed (dominant option). 10. Both ICERs are above the current cost-effectiveness threshold of £20. If the ICER for a given option is higher than the ICER calculated for the previous intervention in ranking. mirtazapine. and excluding from each new incremental analysis all options found to be cost effective in previous analyses. then this strategy is also excluded from further analysis on the basis of extended dominance. based on the results of the deterministic analysis. Escitalopram and venlafaxine both fell slightly in the cost-effective rankings because escitalopram remains under patent and 408 . mirtazapine. Overall. followed by sertraline.172 per QALY for severe depression. venlafaxine. In the economic analysis. The rankings of antidepressants in terms of QALYs in Table 89 were identical for both moderate and severe depression. ICERs are recalculated. citalopram and venlafaxine were ranked as the five best antidepressants in terms of cost effectiveness for both moderate and severe depression. sertraline and citalopram were ranked as the five best antidepressants in terms of efficacy (measured by ORs versus fluoxetine).987 per QALY for moderate depression and £27. The ICERs of escitalopram versus sertraline are £32.4 Results Mirtazapine appears to be the most cost-effective option among those assessed for both moderate and severe depression. Reboxetine was ranked last in both cases. it is possible to rank all antidepressants in terms of cost effectiveness. based on the ORs of fluoxetine versus each antidepressant as reported by Cipriani and colleagues (2009). ICERs are calculated for all pairs of consecutive options. ICERs express the additional cost per additional unit of benefit associated with one treatment option compared with another.Pharmacological interventions further analysis. If mirtazapine is not a suitable treatment option for patients with moderate or severe depression. followed by escitalopram.000 per QALY recommended by NICE (NICE. where the antidepressant drugs have been ranked from the most to the least effective in terms of number of QALYs gained.12. By repeating this process in steps.

615 0.595 0. The analysis demonstrated that overall results were robust with mirtazapine remaining the dominant option for both moderate and severe depression.463 0.620 0.5 Discussion – limitations Given the time constraints involved.438 0.458 0.468 0.148 £1946 £1878 £2.Pharmacological interventions Table 89: Mean costs and QALYs associated with each antidepressant assessed for patients with depression Antidepressant Mean QALYs per person MODERATE DEPRESSION Mirtazapine Escitalopram Venlafaxine Sertraline Citalopram Paroxetine Duloxetine Fluvoxamine Fluoxetine Reboxetine 0.596 0. Table 90 presents the rankings of each antidepressant in terms of both their efficacy and cost effectiveness. The study did not analyse older antidepressants including TCAs and MAOIs. a preliminary economic analysis was undertaken based on the results of the network meta-analysis by Cipriani and colleagues (2009).616 0.596 0. of which two (bupropion and milnacipran) were excluded from the economic analysis in this guideline.102 £1798 £1840 £1908 £2.439 0.403 Mean cost per person MODERATE DEPRESSION £1459 £1597 £1781 £1478 £1522 £1590 £1831 £1629 £1561 £1867 SEVERE DEPRESSION £1781 £1918 £2. which is a limitation in terms of the comprehensiveness of the economic analysis presented here.177 venlafaxine has only recently become available in generic form and its price remains high (although it may be expected to fall substantially).12. 10. The study 409 . The model used to compare the cost effectiveness of combination therapy and antidepressant treatment was adapted for this analysis.598 0.441 0.567 SEVERE DEPRESSION 0.462 0. Sensitivity analysis was undertaken to explore uncertainty around the ORs for efficacy and acceptability estimated in the network meta-analysis by using the upper and lower limits of the 95% credible intervals.438 0.The other three antidepressants are available in generic form and hence much cheaper. The network meta-analysis examined 12 new-generation antidepressants.446 0.602 0.612 0.

A more comprehensive economic analysis would be able to consider costs and outcomes over a longer time horizon. evaluated efficacy regarding response and acceptability in terms of dropouts over the acute phase of depression (8 weeks). Another issue relates to the current and future costs of the antidepressants analysed. Other possible limitations of the study have been highlighted in the clinical review in Section 10. it is 410 . if a patient discontinues their initial antidepressant because of adverse side affects or other factors. although the current price listed in the NHS Drug Tariff (NHS.Pharmacological interventions Table 90: Rankings of each antidepressant in terms of efficacy and cost effectiveness Efficacy (response)* Cost effectiveness MODERATE depression (1) Mirtazapine (2) Escitalopram (3) Venlafaxine (4) Sertraline (5) Citalopram (6) Paroxetine (7) Fluoxetine (8) Duloxetine (9) Fluvoxamine (10) Reboxetine (1) Mirtazapine (2) Sertraline (3) Escitalopram (4) Citalopram (5) Venlafaxine (6) Paroxetine (7) Fluoxetine (8) Fluvoxamine (9) Duloxetine (10) Reboxetine SEVERE depression (1) Mirtazapine (2) Sertraline (3) Escitalopram (4) Citalopram (5) Venlafaxine (6) Paroxetine (7) Fluoxetine (8) Duloxetine (9) Fluvoxamine (10) Reboxetine *Adapted from Cipriani et al. In clinical practice. discontinuation symptoms and social functioning were not investigated in the meta-analyses. As Cipriani and colleagues (2009) acknowledge. ranked according to ORs versus fluoxetine as reference compound. other important outcomes such as side effects.11. (2009). consider issues of drug sequencing or switching (for patients who discontinue initial antidepressant treatment). which is another possible limitation. Similarly. 2009) remains high. it is anticipated that this price will fall further to non-proprietary levels. toxic effects. Venlafaxine has recently been available in generic form and. Business Services Authority. another second-line antidepressant would almost certainly be offered. and give more explicit consideration (captured in estimation of QALYs) of the side effects of different antidepressants as well as impacts on patient mortality (because of side effects or increased suicide risk). The economic analysis did not consider the possibility of switching to secondline antidepressants for patients who discontinue their first-line antidepressant.

13 FROM EVIDENCE TO RECOMMENDATIONS Apart from the review of escitalopram. In addition. A review of the clinical evidence for the new antidepressant drug duloxetine was added. but the drug was found to be no more clinically effective than other antidepressant drugs.12. and the underlying assumptions behind them. is not altered. Ideally. The overall conclusion that antidepressants have largely equal efficacy and that choice should largely depend on side-effect profile.Pharmacological interventions anticipated that escitalopram will shortly be available in generic form. The economic evidence on escitalopram showed it to be more cost effective in comparison with three other antidepressants. No advantage for so-called ‘dual-action’ antidepressants as a class over other drugs was found. Addressing these issues may alter the outcome of the model. However. Another major limitation of this economic model is the inadequate exploration of uncertainty around the results in terms of the assumptions and the clinical efficacy and acceptability data used. the pharmacoeconomic evidence on duloxetine was contradictory and. comprehensive deterministic sensitivity analyses are required. the economic evidence had limitations and these comparisons were considered insufficient to make a specific recommendation for this treatment. it could not be specifically recommended. The updated review of escitalopram showed a small advantage over other antidepressants. is also required. therefore.6 Conclusions The findings of this preliminary economic analysis suggest that mirtazapine might be more cost effective than other antidepressants in the treatment of people with moderate and severe depression and support the findings of Cipriani and colleagues (2009) regarding the clinical superiority of mirtazapine. patient preference and previous experience of treatments. which demonstrates the joint uncertainty between all of the different parameters used in the model. propensity to cause discontinuation symptoms and safety in overdose. therefore. Given the considerable uncertainty around some of the input parameters used in the model. However. 10. but this was not judged to be clinically important over other antidepressants. probabilistic sensitivity analysis. 10. However. this would have necessitated full access to the results (posterior estimates of every iteration) of the network meta-analysis by Cipriani and colleagues (2009). their relative cost effectiveness compared with other antidepressants is likely to be further improved. the recommendations concerning the choice of antidepressants have been updated only to ensure compatibility with the current NICE house style. these economic findings are subject to considerable uncertainty arising from the limitations of the current model and lack of incorporation of data on the relative adverse effects of the drugs in the model. 411 . the reviews of individual drugs undertaken for the previous guideline were not updated and. As the prices of both antidepressants in generic form are likely to be lower in the future.

1 CLINICAL PRACTICE RECOMMENDATIONS Choice of antidepressant147 10.1. side effects and discontinuation symptoms (see Section 11. NICE.1 Discuss antidepressant treatment options with the person with depression.8.14.2 When an antidepressant is to be prescribed. 148Consult Appendix 1 of the BNF for information on drug interactions and ‘Depression in adults with a chronic physical health problem’ (Clinical Guideline 91. it should normally be an SSRI in a generic form because SSRIs are equally effective as other antidepressants and have a favourable risk–benefit ratio. The GDG considered the findings from of the review by Cipriani and colleagues (2009) and developed an economic model based on the review. see also NICE clinical guideline 45 ‘Antenatal and postnatal mental health’. for example. 412 . 2009c). please refer to the BNF and individual drug Summary of Product Characteristics. More information on this topic is provided in the NICE guideline on treating depression in adults with a chronic physical health problem (NICE.7.1. 10. 147For additional considerations on the use of antidepressants and other medications (including the assessment of the relative risks and benefits) for women who may become pregnant. consider prescribing a gastroprotective drug in older people who are taking non-steroidal anti-inflammatory drugs (NSAIDs) or aspirin.2) and potential interactions with concomitant medication or physical health problems148 ● their perception of the efficacy and tolerability of any antidepressants they have previously taken. and these drugs are generally preferred on grounds of cost. especially in older people or in people taking other drugs that have the potential to damage the gastrointestinal mucosa or interfere with clotting.14. An increasing number of newer antidepressants are available as generics.14 10. In particular. covering: ● the choice of antidepressant. Clinicians should also consider the potential for drug interactions when prescribing an antidepressant for people taking concomitant medication.Pharmacological interventions including considering duloxetine and venlafaxine separately. Also take the following into account: ● SSRIs are associated with an increased risk of bleeding. 2009c). The GDG concluded that the analysis was consistent with the findings from the analyses undertaken for this guideline in suggesting some efficacy and tolerability differences between individual antidepressants.14. including any anticipated adverse events. 10. However the size of effect and concerns about potential confounds meant that the findings were not considered sufficiently robust to warrant singling out individual drugs for recommendation. For women in the antenatal and postnatal periods.

in the large naturalistic STAR*D study (Trivedi et al.1. the curve begins to flatten off thereafter. For example. 2006).Pharmacological interventions Fluoxetine.3 When prescribing drugs other than SSRIs. Posternak and Zimmerman (2005). 151Consult the BNF for detailed information. such as phenelzine. duloxetine and TCAs. with a smaller degree of improvement as time goes on. However. and 25% between weeks 1 and 2. should normally be prescribed only by specialist mental health professionals. with the greatest degree of improvement occurring in the first week. ● Dosulepin should not be prescribed. in a meta-analysis of 47 placebocontrolled studies followed up at 6 weeks. contraindications and monitoring requirements for some drugs. 10.15 WHEN TO CHANGE ANTIDEPRESSANT TREATMENT WHEN SYMPTOMS OF DEPRESSION ARE NOT IMPROVING Introduction 10. found that 35% of the improvement occurred between weeks 0 and 1. ● Non-reversible monoamine oxidase inhibitors (MAOIs). 413 . it is important to recognise that although the curve flattens. fluvoxamine and paroxetine are associated with a higher propensity for drug interactions than other SSRIs149. Any wording changes have been made for clarification only.14. ● 10.. ● The specific cautions. 150The evidence for this recommendation has not been updated since the previous guideline. For example: – the potential for higher doses of venlafaxine to exacerbate cardiac arrhythmias and the need to monitor the person’s blood pressure – the possible exacerbation of hypertension with venlafaxine and duloxetine – the potential for postural hypotension and arrhythmias with TCAs – the need for haematological monitoring with mianserin in elderly people151. 149Ibid. some people continue to improve after this and the assessment of the literature is influenced by the duration of follow-up. take the following into account: ● The increased likelihood of the person stopping treatment because of side effects (and the consequent need to increase the dose gradually) with venlafaxine. This is now recognised as incorrect and it has been shown from data from clinical trials that improvement can start immediately.15. ● Paroxetine is associated with a higher incidence of discontinuation symptoms than other SSRIs150.1 Received wisdom has been that antidepressants have a delayed onset of action and that it takes 2 to 4 weeks for them to begin to work.

There is limited but consistent evidence.7 weeks with about 65% of patients responding by 6 weeks. in a systematic review of 41 RCTs. Delaying change in treatment too long could prolong the period of depression if symptoms are not going to respond to current drug/dose. 1999. Increasing the dose too early could lead to patients unnecessarily being maintained on higher than needed doses of antidepressants over a prolonged period of time with associated increased side effects or treatment discontinuation (Bollini et al.. However.Pharmacological interventions which enrolled nearly 2. The rate and degree of improvement also appears to be influenced by the frequency of follow up. 2003. mostly from studies with SSRIs. that increasing the dose after 3 weeks treatment in those not responding ( 50% decrease in rating scale) or remitting (HAMD 9) at this early stage does not improve outcome at 6 weeks (Adli et al. 10.876 patients followed up to 12 weeks. Szegedi et al. Furukawa et al.15. 2005)... 58 to 91% of those responding by 24 weeks had done so by 6 weeks and 79 to 98% responded by 12 weeks. 414 . 2009).. These studies emphasise the importance of the early stages of treatment in response to antidepressants and highlight the role of frequency of monitoring. Depending on the treatment arm. 1995. This is consistent with usual clinical practice. found that weekly assessment between weeks 2 and 6 led to a greater reduction in HAMD score than less frequent assessment in a dose-related manner. This applied to both placebo and drug treatment arms and they estimated that follow-up frequency accounted for about 40% of the placebo response. Switching treatment too early could lead to rejection of an effective treatment. which in the long run will be unhelpful when future treatment options are considered and could lead to a merry-goround of treatment changes.2 Early prediction of eventual response Most studies have found that early improvement in the first 2 weeks (20% or greater improvement) is a good predictor of response by the end of the study (Nierenberg et al. Response was defined as at least a 50% improvement on the HAMD plus at least much improved and not more than mildly ill on the CGI. the mean time to response with citalopram (defined by at least 50% reduction in the Quick Inventory of Depressive Symptomatology-Self Report [QIDS-SR]) was 5. but some patients continuing to respond at 12 weeks and beyond. these are stringent criteria and do not allow clinicians to judge whether altering treatment is beneficial in those not improving at all. Szegedi et al.. A key issue related to this is the optimum time to change treatment. mianserin or placebo. lead to a patient’s loss of faith in treatment. Nierenberg et al.. Posternak and Zimmerman (2007).. Stassen and colleagues (1993) found that the natural variation in assessment makes the minimum reliably detectable improvement in a rating scale in the range of 15 to 25% and most subsequent studies have examined the predictive value of non-improvement using a criterion of 20% or less (these are referred to here as 20% improvers). and increase depression-related morbidity and even mortality. Malt and colleagues (1999) undertook an RCT of 372 primary care patients randomised to sertraline. 2000. 2002b). and treated for 24 weeks.

Higher early NPVs are associated with shorter studies and RCTs and lower NPVs with longer. in other words only 11% would have a stable response as defined. changing treatment in non-responders to the first treatment only resulted in an average increase in response of about 30%. 2009) with a variable follow-up (discharge. Nierenberg et al. They found that 60 to 76% of patients on antidepressants compared with 52% on placebo were 20% improvers at 2 weeks. It is possible to draw only tentative conclusions from these studies. 74 to 83% at 3 weeks and 96 to 99% at 4 weeks (Stassen et al. adding another drug or changing the drug. This can be assessed using the negative predictive value (NPV). A 5-week study found an NPV for 20% improvement on response at 5 weeks of 48 to 54% at 2 weeks... 1993). 2000) and a pooled analysis of 14 escitalopram studies (20% improvement) (Baldwin et al. in the large naturalistic STAR*D study (Rush et al. an open 12-week study of fluoxetine (Quitkin et al. 1995. 77 to 94% at 3 weeks and 82 to 94% at 4 weeks (Szegedi et al. Where this is low.. non-improvement at that time point is not a useful predictor of outcome at endpoint. mean 60 days). using 20% improvement found only a 37% NPV at 2 weeks for response and 43% at 4 weeks.. which is the proportion of non-improvers not going on to achieve response/remission at the last evaluation. Most studies were only 6 weeks in duration.. a reasonable time to consider a change of treatment in these patients would be at 3 to 4 weeks. 2003) using 25% improvement to predict remission (HAMD 8) found an NPV of only 49% at 4 weeks. 2006). A matter for debate is: what is a reasonable value for the NPV that should trigger change of treatment? It is suggested that it lies somewhere between 70 and 80%. The strongest recent case for changing treatment as early as 2 weeks in nonimprovers is Szegedi and colleagues (2009). open studies and possibly more severe patients. in other words. 80 to 82% at 4 weeks and 90 to 93% at 6 weeks. Two 8-week studies of fluoxetine (defining improvement as 20% reduction in one and 30% in another) (Nierenberg et al. It is useful to consider other studies according to the length of follow up. Taking the middle ground with an assessment period of 8 weeks and an NPV based on less than 20% improvement predicting lack of response at 8 weeks. 2003).. 2009) found NPVs of 55 to 64% at 2 weeks. Using a definition of stable response (response at both 4 weeks and subsequently. In patients who have failed 415 .Pharmacological interventions The outcome of concern is the number of non-improvers at each time point who subsequently respond or remit by the end of a certain time frame because this provides some guide as to when changing treatment is likely to improve outcome.. 59 to 69% at 6 weeks and 77% at 8 weeks. Finally. The limitations are the short time frame of most of the studies and the requirement to have responded by 4 weeks. usually 6 weeks). A 6-week study found an NPV defined in the same way as 65 to 72% at 2 weeks. a naturalistic study of 795 inpatients (Henkel et al.. To put this into context. They pooled data from 41 RCTs in which mirtazapine was compared with active comparators or placebo. the overall NPV for those not having a 20% improvement at 2 weeks was 89%. if a non-improver still has more than a 20 to 30% chance of responding then it is probably reasonable to persist longer with treatment before adding to the potential side-effect burden by increasing the dose. In contrast. NPVs for remission (HAMD 8) were higher at 69% and 72% respectively.

Taken together these led to the recommendation that if there is no. provide appropriate information and consider one of the following strategies: ● monitor symptoms closely where side effects are mild and acceptable to the person or ● stop the antidepressant or change to a different antidepressant if the person prefers or ● in discussion with the person.2 If a person with depression develops side effects early in antidepressant treatment. and then at longer intervals if response is good.3 From evidence to recommendations Antidepressant studies examining the onset of improvement in relation to response or remission at the end of the study vary in their findings according to the exact methodology used. which was thought by the GDG to present a realistic clinical situation. See them regularly thereafter. 10. consider short-term concomitant treatment with a benzodiazepine if anxiety. agitation and/or insomnia are problematic (except in people with chronic symptoms of anxiety). when evaluated at 2 weeks these patients had about a 40% chance of achieving a response at 8 weeks falling to a 20% chance if they had failed to improve by 4 weeks.15. the GDG noted that there is some evidence that a higher frequency of assessment between weeks 2 and 6 is associated with a better outcome. longer trials of treatment may be warranted before making changes.4 Clinical practice recommendations 10.Pharmacological interventions previous trials of treatment. for example. and in more severely ill patients. Any wording changes have been made for clarification only. The rate of improvement after 6 to 8 weeks of treatment is slower and only a minority of non-responders at this stage will go on to have an adequate response over the next 6 to 18 weeks. 10.15. this 152The evidence for this recommendation has not been updated since the previous NICE guideline.4.1 For people started on antidepressants who are not considered to be at increased risk of suicide. response. or barely detectable. and using less than a 20% improvement on the HAMD score to indicate patients with a lack of. patients should be followed weekly and consideration given to changing treatment at 3 to 4 weeks. at intervals of 2 to 4 weeks in the first 3 months.4. In addition. Taking studies evaluating response over an 8-week time frame.152 10. normally see them after 2 weeks. or barely any detectable improvement at 2 to 4 weeks.15. consideration should be given to changing treatment at that stage. 416 .15. Patients who are improving should have their improvement monitored and if there has been insufficient response at 6 weeks in the absence of a continuing trajectory of improvement.

15.15.Pharmacological interventions should usually be for no longer than 2 weeks in order to prevent the development of dependence.3. check that the drug has been taken regularly and in the prescribed dose. 10.15. 10.3. continue treatment for another 2 to 4 weeks.4. Consider switching to another antidepressant as described in 12. increase the level of support (for example.3 If the person’s depression shows no improvement after 2 to 4 weeks with the first antidepressant. 10. 417 .4 If response is absent or minimal after 3 to 4 weeks of treatment with a therapeutic dose of an antidepressant.4.16 if there are side effects or if the person prefers.4.5 If the person’s depression shows some improvement by 4 weeks.16 if: ● response is still not adequate or ● there are side effects or ● the person prefers to change treatment. by weekly face-to-face or telephone contact) and consider: ● increasing the dose in line with the Summary of Product Characteristics if there are no significant side effects or ● switching to another antidepressant as described in Section 12.

This chapter updates the reviews on the effect of sex on antidepressant choice.3) ● the pharmacological management of depression with psychotic symptoms (Section 11.10). cardiotoxicity of antidepressants. the GDG felt that the previous recommendations should be removed since there was no reason why treatment for people whose depression had atypical features should not follow that for those with major depression. In addition. which covers many issues relevant to older people with depression (NICE. 2010). antidepressant discontinuation symptoms.4) ● the pharmacological management of atypical depression (Section 11.Factors influencing choice of antidepressants 11 FACTORS INFLUENCING CHOICE OF ANTIDEPRESSANTS 11. However.7) ● antidepressant discontinuation symptoms (Section 11.2) ● the effect of sex on antidepressant choice (Section 11.9) ● depression. 11. 418 .8) ● the cardiotoxicity of antidepressants (Section 11. a separate guideline has been developed specifically for depression in adults with a chronic physical health problem. The section on depression with psychotic sypmtoms was not updated and the recommendations were left unchanged. including: ● the pharmacological management of depression in older adults (Section 11. since the previous guideline. NCCMH. 2009c. The review of atypical depression was also not updated. and antidepressants and suicide. It includes a new review of treatments for depression with a seasonal pattern because this diagnosis was added to the scope of the updated guideline.5) ● the physical and pharmacological management of depression with a seasonal pattern (Section 11.2 THE PHARMACOLOGICAL MANAGEMENT OF DEPRESSION IN OLDER ADULTS The following sections on the pharmacological management of depression in older adults marked by asterisks (**_**) are from the previous guideline and have not been updated except for style and minor clarification. antidepressants and suicide (Section 11. The review of low-dose versus high-dose TCAs was not updated.1 INTRODUCTION While the previous chapter reviewed the relative efficacy of different antidepressants.6) ● dosage issues for tricyclic antidepressants (Section 11. The review of the pharmacological management of depression in older adults was not updated because there were little new data in older adults to indicate that the existing recommendations should be amended. this chapter looks at factors that may affect the choice of antidepressant.

. phenelzine.. which also contains a list of excluded studies with reasons for exclusions. The efficacy of antidepressants in older adults has been summarised in a Cochrane systematic review (Wilson et al. each study considered for review is referred to by a ‘study ID’ made up of first author and publication date (unless a study is in press or only submitted for publication. 1997). TCAs as a group. 1997) and sub-optimal treatment (Iliffe et al. to start low and increase slowly but it is clear that much more research involving older patients with depression is required on this and other points. Most depression in older adults is treated in primary care (Plummer et al. References for studies from the previous guideline are in Appendix 18 and references for studies for the update are in Appendix 17c.. 2001).2.2 Use of individual antidepressants in the treatment of depression in older adults Studies considered153. 153Details 419 . 154Here and elsewhere in the guideline. sensory deficits and other disabilities and. Untreated it shortens life and increases healthcare costs. therefore. It was possible to review the following pharmacological strategies for the treatment of depression in older adults: ● use of individual antidepressants (amitriptyline. therefore.. studies were also available for reboxetine but. 1999). since they might substantially affect function in a vulnerable individual. There is some evidence that older people take longer to recover than younger adults and adverse events need to be carefully monitored for. The maxim is.2.Factors influencing choice of antidepressants 11. when first author only is used). as well as adding to disability from medical illnesses. In this population the monitoring of self-harm is particularly important. 1997) but there is evidence of poor detection (Plummer et al.154 This review brings together studies from other reviews undertaken for this guideline where more than 80% of study participants were aged 65 years and over. A separate of standard search strings used in all searches are in Appendix 8. venlafaxine) and St John’s wort. 11.. medication needs to be carefully monitored in these groups. Older adults are at risk of co-existing physical disorders. mirtazapine.1 Introduction **Depression is the most common mental health problem of later life affecting approximately 15% of older people (Beekman et al. There are a variety of potential differences in older adults in terms of absorption and metabolism of drugs and increased potential for interaction with other drugs. it is not reviewed ● augmentation of an antidepressant with lithium ● strategies for relapse prevention. since this drug is not licensed for the treatment of depression in older adults.. 1991). It is also very important to educate the patient and caregivers about depression and involve them in treatment decisions. Information about each study along with an assessment of methodological quality is in Appendix 17c. Study IDs in title case refer to studies included in the previous guideline and study IDs in capital letters refer to studies found and included in this guideline update. SSRIs. and is the leading cause of suicide among older people (Lebowitz et al.

N 602. studies undertaken with elderly populations using drugs not reviewed for this guideline are not included. SMD –0.02. Rahman1991. two from venlafaxine and one each from mirtazapine. eight as outpatient and one as primary care. N 202. Ten studies were sourced from the review of SSRIs. –0. Efficacy data were available from up to 1. 155The forest plots can be found in Appendix 19c. N 254.27) ● between SSRIs and alternative antidepressants (K 8. 420 . N 149.19) ● between St John’s wort and fluoxetine (K 1. All included studies were published between 1985 and 2002. 95% CI. –0. Feighner1985a. There is insufficient evidence to determine if there is a clinically important difference in older adults on increasing the likelihood of achieving a 50% reduction in symptoms of depression between: ● amitriptyline and paroxetine ● venlafaxine and TCAs ● alternative antidepressants and TCAs ● St John’s wort and fluoxetine ● mirtazapine and paroxetine. Two were classified as inpatient. –0. Geretsegger95.00. SMD 0.620 patients. –0. Studies ranged from 5 to 8 weeks long. SMD –0. participants were either from mixed sources or it was not possible determine the source. In four. In all. Pelicier1993.01. and tolerability data from up to 1.36 to 0. Schatzberg02. Clinical evidence statements155 Effect of treatment on efficacy There is evidence suggesting that there is no clinically important difference on reducing symptoms of depression in older adults: ● between amitriptyline and paroxetine (K 2. phenelzine and St John’s wort. Hutchinson92. 95% CI. 95% CI. Phanjoo1991.1. Studies were included provided the mean dose achieved was at least half the ‘standard’ adult dose. SMD –0.04.26 to 0. Smeraldi1998).29) ● between alternative antidepressants and TCAs (K 6. SMD –0. Dorman1992. 95% CI.Factors influencing choice of antidepressants systematic search of the literature was not undertaken and.19 to 0.46 to 0. SMD 0.28) ● between mirtazapine and paroxetine (K 1. N 443. GeorgotaS86. LaPia1992. –0.15) ● between venlafaxine and TCAs (K 2.17 to 0. 95% CI. N 126. Mahapatra1997. 15 studies from other reviews of individual antidepressants enrolled participants who were at least 60 years of age (Cohn1990. therefore. Harrer99.37 to 0. Guillibert89. –0.13). 95% CI.083 patients.12.

0. 95% CI. N 254. There is evidence suggesting that there is no clinically important difference between alternative antidepressants and TCAs on reducing the likelihood of older adults reporting adverse effects (K 7. N 254. N 95. N 60. N 1115. RR 0. –0.97.09) ● mirtazapine and paroxetine (K 1.55.87 to 1.73 to 1. 95% CI.33). 0.02).79 to 1. There is insufficient evidence to determine if there is a clinically important difference in older adults on increasing the likelihood of achieving remission: ● between phenelzine and nortriptyline ● alternative antidepressants and TCAs.12) ● SSRIs and alternative antidepressants (K 10.83 to 1. RR 0.87.95.96. RR 0. SMD –0. 95% CI. N 254. 421 . 95% CI. There is insufficient evidence to determine any difference on any efficacy measure in older outpatients or patients in primary care.48 to 0.57. RR 0. 95% CI.97.82 to 1. RR 0.03). N 1058.86 to 1. 0. 95% CI. 0.09). 95% CI. RR 0. 0.34 to 0. Acceptability and tolerability of treatment There is some evidence suggesting that there is a clinically important difference favouring mirtazapine over paroxetine on reducing the likelihood of older adults leaving treatment early due to side effects (K 1. RR 0. N 422.13). N 90. 95% CI.86). Effect of setting on treatment efficacy and tolerability There is evidence suggesting that there is no clinically important difference between SSRIs and TCAs on reducing symptoms of depression in older inpatients (K 2. N 1154. There is evidence suggesting that there is no clinically important difference on reducing the likelihood of older adults reporting adverse events between: ● SSRIs and alternative antidepressants (K 8.7 to 1. 0.13) ● alternative antidepressants and TCAs (K 10. There is evidence suggesting that there is no clinically important difference on reducing the likelihood of older adults leaving treatment early between: ● amitriptyline and SSRIs (K 3. 0.89.23). 0. 95% CI. RR 1.05) ● phenelzine and nortriptyline (K 1. There is evidence suggesting that there is no clinically important difference between SSRIs and alternative antidepressants on reducing the likelihood of older adults leaving treatment early due to side effects (K 10. 95% CI. N 717.97. There is insufficient evidence to determine if there is a clinically important difference between other drug comparisons on other tolerability measures. 0. There is some evidence suggesting that there is a clinically important difference favouring paroxetine over amitriptyline on reducing the likelihood of older adults in primary care reporting adverse effects (K 1. 0. 95% CI.Factors influencing choice of antidepressants There is evidence suggesting that there is no clinically important difference between mirtazapine and paroxetine on increasing the likelihood of achieving remission in older adults (K 1.81 to 1. RR 0.85 to 1. N 581.35 to 0. 95% CI.94).07. 0. RR 0.89. RR 0.

Information about each study along with an assessment of methodological quality is in Appendix 17c. 11. RR 5. Information ab